Solarium Vitamin D Rã¶Hren

Abstract

Following a request from the European Commission, the EFSA Panel on Dietetic Products, Nutrition and Allergies (NDA) derived dietary reference values (DRVs) for vitamin D. The Panel considers that serum 25(OH)D concentration, which reflects the amount of vitamin D attained from both cutaneous synthesis and dietary sources, can be used as a biomarker of vitamin D status in adult and children populations. The Panel notes that the evidence on the relationship between serum 25(OH)D concentration and musculoskeletal health outcomes in adults, infants and children, and adverse pregnancy-related health outcomes, is widely variable. The Panel considers that Average Requirements and Population Reference Intakes for vitamin D cannot be derived, and therefore defines adequate intakes (AIs), for all population groups. Taking into account the overall evidence and uncertainties, the Panel considers that a serum 25(OH)D concentration of 50 nmol/L is a suitable target value for all population groups, in view of setting the AIs. For adults, an AI for vitamin D is set at 15 μg/day, based on a meta-regression analysis and considering that, at this intake, the majority of the population will achieve a serum 25(OH)D concentration near or above the target of 50 nmol/L. For children aged 1–17 years, an AI for vitamin D is set at 15 μg/day, based on the meta-regression analysis. For infants aged 7–11 months, an AI for vitamin D is set at 10 μg/day, based on trials in infants. For pregnant and lactating women, the Panel sets the same AI as for non-pregnant non-lactating women, i.e. 15 μg/day. The Panel underlines that the meta-regression was done on data collected under conditions of assumed minimal cutaneous vitamin D synthesis. In the presence of cutaneous vitamin D synthesis, the requirement for dietary vitamin D is lower or may even be zero.

Summary

Following a request from the European Commission, the EFSA Panel on Dietetic Products, Nutrition and Allergies (NDA) was asked to deliver a Scientific Opinion on dietary reference values (DRV) for the European population, including vitamin D.

Vitamin D is the generic term for ergocalciferol (vitamin D₂) and cholecalciferol (vitamin D₃), which are formed from their respective provitamins, ergosterol and 7-dehydrocholesterol, following a two step-reaction involving ultraviolet-B (UV-B) irradiation and subsequent thermal isomerisation. Vitamin D₂ and vitamin D₃ are fat-soluble and present in foods and dietary supplements. Vitamin D₃ is also synthesised endogenously in the skin following exposure to UV-B irradiation.

During summer months, or following exposure to artificial UV-B irradiation, the synthesis of vitamin D₃ in the skin may be the main source of vitamin D. Dietary intake of vitamin D is essential in case endogenous synthesis, due to insufficient UV-B exposure, is lacking or insufficient. Factors affecting the synthesis of vitamin D₃ in the skin include latitude, season, ozone layer and clouds (absorbing UV-B irradiation), surface characteristics (reflecting UV-B irradiation), time spent outdoors, use of sunscreen, clothing, skin colour and age. The Panel notes that sun exposure may contribute a considerable and varying amount of vitamin D available to the body and therefore considers that the association between vitamin D intake and status, for the purpose of deriving DRVs for vitamin D, should be assessed under conditions of minimal endogenous vitamin D synthesis. Vitamin D from dietary sources is absorbed throughout the small intestine. The Panel considers that the average vitamin D absorption from a usual diet is about 80% and limited data are available on the effect of the food or supplement matrix on absorption of vitamin D (vitamin D₂ or vitamin D₃).

In the body, within hours of ingestion or synthesis in the skin, vitamin D is either converted into its biologically active metabolite 1,25(OH)₂D or delivered to the storage tissues (as either vitamin D or its metabolites). The first step of the conversion occurs in the liver, where vitamin D is hydroxylated to 25(OH)D, while the second step occurs primarily in the kidneys, where 25(OH)D is hydroxylated to 1,25(OH)₂D. Vitamin D, 1,25(OH)₂D and 25(OH)D are transported in the blood bound mainly to the vitamin D-binding protein (DBP). Of the two metabolites of vitamin D, 25(OH)D is the major circulating form, with a longer mean half-life, of about 13–15 days. 25(OH)D is taken up from the blood into many tissues, including in the adipose tissue, muscle and liver for storage.

After its release from DBP to tissues, 1,25(OH)₂D exerts, in association with the intracellular vitamin D receptor (VDR), important biological functions throughout the body. In the intestine, it binds to VDR to facilitate calcium and phosphorus absorption. In the kidney, it stimulates the parathyroid hormone (PTH)-dependent tubular reabsorption of calcium. In the bone, PTH and 1,25(OH)₂D interact to activate the osteoclasts responsible for bone resorption. In addition, 1,25(OH)₂D suppresses the PTH gene expression, inhibits proliferation of parathyroid cells, and is involved in cell differentiation and antiproliferative actions in various cell types. Both 25(OH)D and 1,25(OH)₂D are catabolised before elimination and the main route of excretion is via the faeces.

Vitamin D deficiency leads to impaired mineralisation of bone due to an inefficient absorption of dietary calcium and phosphorus, and is associated with an increase in PTH serum concentration. Clinical symptoms of vitamin D deficiency manifest as rickets in children, and osteomalacia in adults.

The Panel reviewed possible biomarkers of vitamin D intake and/or status, namely serum concentration of 25(OH)D, free 25(OH)D, 1,25(OH)₂D and PTH concentration, markers of bone formation and bone turnover. In spite of the high variability in 25(OH)D measurements obtained with different analytical methods, the Panel concludes that serum 25(OH)D concentration, which reflects the amount of vitamin D attained from both cutaneous synthesis and dietary sources, can be used as a biomarker of vitamin D status in adult and children populations. Serum 25(OH)D concentration can also be used as a biomarker of vitamin D intake in a population with low exposure to UV-B irradiation.

In consideration of the various biological functions of 1,25(OH)₂D, the Panel assessed the available evidence on the relationship between serum 25(OH)D concentration and several health outcomes, to evaluate whether they might inform the setting of DRVs for vitamin D. The Panel first considered the available evidence on serum 25(OH)D concentration and musculoskeletal health outcomes, i.e. bone mineral density (BMD)/bone mineral content (BMC) and calcium absorption in adults and infants/children, risk of osteomalacia, fracture risk, risk of falls/falling, muscle strength/muscle function/physical performance in adults, and risk of rickets in infants/children. The Panel then reviewed data on the relationship between maternal serum 25(OH)D concentration and health outcomes in pregnancy (risk of pre-eclampsia, of small for gestational age and of preterm birth, and indicators of bone health in infants) and lactation. The Panel took as a starting point the results of the literature search and the conclusions from the most recent report on DRVs for vitamin D by the Institute of Medicine (IOM) that was based on two systematic reviews. The Panel also considered an update of one of these two systematic reviews, as well as two recent reports from DRV-setting bodies. The Panel undertook a separate literature search to identify primary intervention and prospective observational studies in healthy subjects (infants, children and adults, including free-living older adults) that were published after the IOM report until March 2015. As a second step, the Panel considered available evidence on several other non-musculoskeletal health outcomes (e.g. cancer or cardiovascular diseases), based on the reports and reviews mentioned above without undertaking a specific literature search of primary studies. The Panel considers that the available evidence on serum 25(OH)D concentration and musculoskeletal health outcomes and pregnancy-related health outcomes is suitable to set DRVs for vitamin D for (healthy) adults, infants, children, and pregnant women, respectively. However, the Panel considers that there is no evidence for a relationship between serum 25(OH)D concentration and health outcomes of lactating women that may be used to set a DRV for vitamin D, and that the available evidence on non-musculoskeletal health outcomes is insufficient to be used as criterion for setting DRVs for vitamin D.

The Panel notes that data on the relationship between serum 25(OH)D concentration and adverse musculoskeletal or pregnancy-related health outcomes are widely variable. However, taking into account the overall evidence and uncertainties, the Panel considers that, for adults, infants and children, there is evidence for an increased risk of adverse musculoskeletal health outcomes at serum 25(OH)D concentrations below 50 nmol/L. The Panel also considers that there is evidence for an increased risk of adverse pregnancy-related health outcomes at serum 25(OH)D concentrations below 50 nmol/L.

The Panel assessed the available evidence on the relationship between vitamin D intake and musculoskeletal health outcomes to evaluate whether they might inform the setting of DRVs for vitamin D. The Panel notes that these studies usually do not provide information on the habitual dietary intake of vitamin D, and the extent to which cutaneous vitamin D synthesis has contributed to the vitamin D supply (and thus may have confounded the relationship between vitamin D intake and the reported health outcomes) is not known. The Panel therefore concludes that these studies are not useful as such for setting DRVs for vitamin D, and may only be used to support the outcome of the characterisation of the vitamin D intake-status relationship undertaken by the Panel under conditions of assumed minimal endogenous vitamin D synthesis.

The Panel concludes that a serum 25(OH)D concentration of 50 nmol/L is a suitable target value to set the DRVs for vitamin D, for all age and sex groups (healthy adults, infants, children, pregnant and lactating women). For setting DRVs for vitamin D, the Panel considers the dietary intake of vitamin D necessary to achieve this serum 25(OH)D concentration. As for other nutrients, DRVs for vitamin D are set assuming that intakes of interacting nutrients, such as calcium, are adequate.

EFSA undertook a meta-regression analysis of the relationship between serum 25(OH)D concentration and total vitamin D intake (habitual diet, and fortified foods or supplements using vitamin D₃). Randomised trials conducted in a period of assumed minimal endogenous vitamin D synthesis were identified through a comprehensive literature search and a review undertaken for EFSA by an external contractor (Brouwer-Brolsma et al., 2016). The analysis was performed using summary data from 83 trial arms (35 studies), of which nine were on children (four trials, age range: 2–17 years) and the other arms were on adults (mean age between 22 and 86 years, excluding pregnant or lactating women). Data were extracted for each arm of the individual trials. The meta-regression analysis resulted in two predictive equations of achieved serum 25(OH)D concentrations: one derived from an unadjusted model (including only the natural log of the total intake) and one derived from a model including the natural log of the total intake and adjusted for a number of relevant factors (baseline serum 25(OH)D concentration, latitude, study start year, type of analytical method applied to assess serum 25(OH)D, assessment of compliance) set at their mean values.

The Panel considers that the available evidence does not allow the setting of average requirements (ARs) and population reference intakes (PRIs), and therefore defines adequate intakes (AIs) instead, for all population groups.

For adults, the Panel sets an AI for vitamin D at 15 μg/day. This is based on the adjusted model of the meta-regression analysis, and considering that, at this intake, the majority of the adult population will achieve a serum 25(OH)D concentration near or above the target of 50 nmol/L.

For children aged 1–17 years, the Panel sets an AI for vitamin D for all children at 15 μg/day. This is based on the adjusted model of the meta-regression analysis on all trials (adults and children) as well as on a stratified analysis by age group (adults versus children).

For infants aged 7–11 months, the Panel sets an AI for vitamin D at 10 μg/day, considering four recent trials on the effect of vitamin D supplementation on serum 25(OH)D concentration in (mostly) breastfed infants.

For pregnant and lactating women, the Panel considers that the AI is the same as for non-pregnant non-lactating women, i.e. 15 μg/day.

The Panel underlines that the meta-regression analysis on adults and children was done on data collected under conditions of assumed minimal cutaneous vitamin D synthesis. In the presence of cutaneous vitamin D synthesis, the requirement for dietary vitamin D is lower or may even be zero.

Background as provided by the European Commission

The scientific advice on nutrient intakes is important as the basis of Community action in the field of nutrition, for example such advice has in the past been used as the basis of nutrition labelling. The Scientific Committee for Food (SCF) report on nutrient and energy intakes for the European Community dates from 1993. There is a need to review and if necessary to update these earlier recommendations to ensure that the Community action in the area of nutrition is underpinned by the latest scientific advice.

In 1993, the SCF adopted an opinion on nutrient and energy intakes for the European Community.1 The report provided Reference Intakes for energy, certain macronutrients and micronutrients, but it did not include certain substances of physiological importance, for example dietary fibre.

Since then new scientific data have become available for some of the nutrients, and scientific advisory bodies in many European Union (EU) Member States and in the United States have reported on recommended dietary intakes. For a number of nutrients, these newly established (national) recommendations differ from the reference intakes in the SCF (1993) report. Although there is considerable consensus between these newly derived (national) recommendations, differing opinions remain on some of the recommendations. Therefore, there is a need to review the existing EU Reference Intakes in the light of new scientific evidence, and taking into account the more recently reported national recommendations. There is also a need to include dietary components that were not covered in the SCF opinion of 1993, such as dietary fibre, and to consider whether it might be appropriate to establish reference intakes for other (essential) substances with a physiological effect.

In this context, the European Food Safety Authority (EFSA) is requested to consider the existing population reference intakes for energy, micro- and macronutrients and certain other dietary components, to review and complete the SCF recommendations, in the light of new evidence, and in addition advise on a population reference intake for dietary fibre.

For communication of nutrition and healthy eating messages to the public, it is generally more appropriate to express recommendations for the intake of individual nutrients or substances in food-based terms. In this context, EFSA is asked to provide assistance on the translation of nutrient based recommendations for a healthy diet into food based recommendations intended for the population as a whole.

Terms of Reference as provided by the European Commission

In accordance with Article 29 (1)(a) and Article 31 of Regulation (EC) No. 178/20022, the Commission requests EFSA to review the existing advice of the Scientific Committee for Food on population reference intakes for energy, nutrients and other substances with a nutritional or physiological effect in the context of a balanced diet which, when part of an overall healthy lifestyle, contribute to good health through optimal nutrition.

In the first instance, EFSA is asked to provide advice on energy, macronutrients and dietary fibre. Specifically advice is requested on the following dietary components:

carbohydrates, including sugars;
fats, including saturated fatty acids, polyunsaturated fatty acids and monounsaturated fatty acids, trans-fatty acids;
protein;
dietary fibre.

Following on from the first part of the task, EFSA is asked to advise on population reference intakes of micronutrients in the diet and, if considered appropriate, other essential substances with a nutritional or physiological effect in the context of a balanced diet which, when part of an overall healthy lifestyle, contribute to good health through optimal nutrition.

Finally, EFSA is asked to provide guidance on the translation of nutrient-based dietary advice into guidance, intended for the European population as a whole, on the contribution of different foods or categories of foods to an overall diet that would help to maintain good health through optimal nutrition (food-based dietary guidelines).

Assessment

1 Introduction

In 1993, the Scientific Committee for Food (SCF) adopted an opinion on nutrient and energy intakes for the European Community and derived for vitamin D acceptable ranges of intakes for adults aged 18–64 years according to the amount of endogenous synthesis of vitamin D while one single value was set for adults aged ≥ 65 years (SCF, 1993). Acceptable ranges of intakes were also set for infants aged 6–11 months, and children aged 4–10 and 11–17 years, according to the amount of endogenous vitamin D synthesis, while a single reference value for the age range 1–3 years was selected. The same reference value was proposed for pregnancy and for lactation.

In the present Opinion, vitamin D intake is expressed in μg and concentrations in blood are expressed in nmol/L.3

2 Definition/category

2.1 Chemistry

Vitamin D is the generic term for ergocalciferol (vitamin D₂) and cholecalciferol (vitamin D₃), which are formed from their respective provitamins ergosterol and 7-dehydrocholesterol (7-DHC) involving ultraviolet-B (UV-B) irradiation that opens the B-ring of the molecules, and subsequent thermal isomerisation (Figure 1). Vitamin D₂ differs from vitamin D₃ in the side chain where it has a double bond between C22 and C23 and an additional methyl group on C24 (Binkley and Lensmeyer, 2010). The molecular masses of ergocalciferol and cholecalciferol are 396.65 and 384.64 g/mol, respectively. In this assessment, the term vitamin D refers to both vitamin D₃ and vitamin D₂ unless the specific form is indicated. Analytical methods for the quantification of vitamin D in serum are discussed in Section 2.4.1.

Vitamins D₂ (ergocalciferol) and D₃ (cholecalciferol) with their respective provitamins. Based on data from Norman (2012)

2.2 Function of vitamin D

2.2.1 Biochemical functions

In the body, vitamin D₂ and D₃ are converted to the main circulating form, 25-hydroxyvitamin D (25(OH)D₂ or 25(OH)D₃ termed calcidiols). It can be transformed into the biologically active metabolites 1,25-dihydroxy-ergocalciferol (1,25(OH)₂D₂) or 1,25-dihydroxy-cholecalciferol (1,25(OH)₂D₃) called calcitriols (Section 2.3.6 and Figure 2). The term 25(OH)D refers to both 25(OH)D₂ and 25(OH)D₃ and 1,25(OH)₂D refers to both 1,25(OH)₂D₃ and 1,25(OH)₂D₂ unless the specific form is indicated.

Metabolism of vitamin D. Based on data from Holick (2006)

The principal function of the biologically active metabolite 1,25(OH)₂D is to maintain calcium and phosphorus homeostasis in the circulation, together with parathyroid hormone (PTH) and fibroblast growth factor (FGF-23) (EFSA NDA Panel, 2012a; Jones, 2013). If the serum ionised calcium concentration falls below a normal concentration of about 1.1–1.4 mmol/L, a cascade of events occurs to restore and maintain it within the range required for normal cellular and tissue functions (Mundy and Guise, 1999; Weaver and Heaney, 2006; Ajibade et al., 2010; EFSA NDA Panel, 2015b). The main target tissues of 1,25(OH)₂D are the intestine, the kidneys and the bone (Figure 2, Section 2.3.6). In the intestine, 1,25(OH)₂D binds to the vitamin D receptor (VDR) to facilitate calcium and phosphorus absorption by active transport. In the kidneys, 1,25(OH)₂D stimulates the tubular reabsorption of calcium dependent on PTH that increases the production of 1,25(OH)₂D from 25(OH)D in the proximal tubule (Holt and Wysolmerski, 2011). 1,25(OH)₂D also downregulates the activity of the enzyme 1α-hydroxylase (CYP27B1), which is responsible for the conversion of 25(OH)D to 1,25(OH)₂D in the kidney. In the bone, PTH and 1,25(OH)₂D interact to activate the osteoclasts responsible for bone resorption. Osteoclasts then release hydrochloric acid and hydrolytic enzymes to dissolve the bone matrix and thereby release calcium and phosphorus into the circulation (Holick, 2006, 2007).

The metabolite 1,25(OH)₂D is also important in other tissues (Bouillon et al., 2008; EFSA NDA Panel, 2012a; Jones, 2014) that have VDRs as well as the 1α-hydroxylase to convert 25(OH)D into 1,25(OH)₂D (Holick, 2007). For example, the parathyroid cells express the VDR and the 1α-hydroxylase, which allows the local formation of 1,25(OH)₂D. 1,25(OH)₂D suppresses the expression of the gene encoding PTH and among other actions, inhibits proliferation of parathyroid cells (Bienaime et al., 2011) (Figure 2).

Other functions of 1,25(OH)₂D include cell differentiation and antiproliferative actions in various cell types, such as bone marrow (osteoclast precursors and lymphocytes), cells belonging to the immune system, skin, breast and prostate epithelial cells, muscle and intestine (Norman, 2008, 2012; Jones, 2014).

2.2.2 Health consequences of deficiency and excess

2.2.2.1 Deficiency

Clinical symptoms of vitamin D deficiency manifest as rickets in children and osteomalacia in adults (Sections 5.1.1, 5.1.2.1.2, 5.1.2.2.2). Both are caused by the impaired mineralisation of bone due to an inefficient absorption of dietary calcium and phosphorus, and both are associated with an increase in serum PTH concentration to prevent hypocalcaemia (Holick, 2006; Holick et al., 2012).

Rickets is characterised by a triad of clinical symptoms: skeletal changes (with deformities, craniotabes, growth retardation), radiologic changes (widening of the metaphyseal plates, decreased mineralisation, deformities) and increases in bone alkaline phosphatase (ALP) activity in serum (Wharton and Bishop, 2003). Depending on the severity and duration of vitamin D deficiency, initial hypocalcaemia progresses to normocalcaemia and hypophosphatemia, because of increased PTH secretion and, finally to combined hypocalcaemia and hypophosphatemia when calcium can no longer be released from bone. Osteomalacia is characterised by increased bone resorption and suppression of new bone mineralisation (Lips, 2006), and serum calcium concentration is often normal (2.25–2.6 mmol/L) despite the undermineralisation of bone. The clinical symptoms of vitamin D deficiency in adults are less pronounced than in children, and may include diffuse pain in muscles and bone and specific fractures. Muscle pain and weakness (myopathy) that accompany the skeletal symptoms in older adults may contribute to poor physical performance, increased risk of falls/falling and a higher risk of bone fractures.

Prolonged vitamin D insufficiency may lead to low bone mineral density (BMD) and may dispose older subjects, particularly post-menopausal women, for osteoporosis, a situation characterised by a reduction in bone mass, reduced bone quality and an increased risk of bone fracture, predominantly in the forearm, vertebrae, and hip (Heaney et al., 2000; Gaugris et al., 2005; Holick, 2007; Avenell et al., 2014).

2.2.2.2 Excess

Following ingestion of pharmacological doses (e.g. 125–1,000 μg/day) of vitamin D over a period of at least 1 month, the concentration of serum 25(OH)D increases, while that of 1,25(OH)₂D is unchanged or even reduced (EFSA NDA Panel, 2012a; Jones, 2014). High serum 25(OH)D concentrations (> 220 nmol/L) may lead to hypercalcaemia, which may eventually lead to soft tissue calcification and resultant renal and cardiovascular damage (Vieth, 1999; Zittermann and Koerfer, 2008).

In revising the tolerable upper intake Levels (ULs) for vitamin D (EFSA NDA Panel, 2012a), data on possible associations between vitamin D intake or serum 25(OH)D concentration and adverse long-term health outcomes were considered. However, no studies reported on associations between vitamin D intake and increased risk for adverse long-term health outcomes. Studies reporting on an association between serum 25(OH)D concentration and all-cause mortality or cancer were inconsistent. For adults, hypercalcaemia was selected as the indicator of hypervitaminosis D or vitamin D toxicity (EFSA NDA Panel, 2012a). Two studies administered doses between 234 and 275 μg/day vitamin D₃ in men without reported hypercalcaemia (Barger-Lux et al., 1998; Heaney et al., 2003b), and a No Observed Adverse Effect Level (NOAEL) of 250 μg/day was established (Hathcock et al., 2007). Taking into account uncertainties associated with these two studies, the UL for adults was set at 100 μg/day. Two studies in pregnant and lactating women, both using doses of vitamin D₂ and D₃ up to 100 μg/day for several weeks to months, did not report adverse effects for either mothers or their offspring (Hollis and Wagner, 2004a; Hollis et al., 2011). Thus, the UL of 100 μg/day applies to all adults, including pregnant and lactating women (EFSA NDA Panel, 2012a).

There is a paucity of data on high vitamin D intakes in children and adolescents. Considering phases of rapid bone formation and growth and the unlikelihood that this age group has a lower tolerance for vitamin D compared to adults, the UL was set at 100 μg/day for ages 11–17 years (EFSA NDA Panel, 2012a). The same consideration applied also to children aged 1–10 years, but taking into account their smaller body size, a UL of 50 μg/day was selected (EFSA NDA Panel, 2012a).

For infants, data relating high vitamin D intakes to impaired growth and hypercalcaemia (Jeans and Stearns, 1938; Fomon et al., 1966; Ala-Houhala, 1985; Vervel et al., 1997; Hyppönen et al., 2011) were used as indicators as in the previous risk assessment by the SCF to set the UL at 25 μg/day (SCF, 2002a). The Panel retained the UL of 25 μg/day and noted that no long-term studies on hypercalcaemia were available (EFSA NDA Panel, 2012a).

The Panel notes that two randomised controlled trials (RCTs) have been published after the assessment of the UL by the EFSA NDA Panel (2012a). In both RCTs, infants received vitamin D₃ supplementation doses ranging between 10 and 40 μg/day, from age 2 weeks to age 3 months (Holmlund-Suila et al., 2012) or from age 1 month to age 12 months (Gallo et al., 2013), with concomitant increases in mean serum 25(OH)D concentrations (Section 5.1.2.2.1). In the shorter term study (Holmlund-Suila et al., 2012), hypercalcaemia or hypercalciuria did not occur at any dose of vitamin D₃ supplemented. In the longer term study (Gallo et al., 2013), the dose of 40 μg/day was discontinued prematurely because of elevated serum 25(OH)D concentrations above 250 nmol/L, a criterion a priori chosen by the authors to indicate hypervitaminosis D.

2.3 Physiology and metabolism

2.3.1 Cutaneous synthesis of vitamin D

Vitamin D₃ is synthesised in the skin from 7-DHC following exposure to UV-B irradiation, which, by opening the B-ring, leads to the formation of previtamin D₃ in the upper layers of the skin that, immediately after its formation, thermally isomerises to vitamin D₃ in the lower layers of the skin (Figure 1) (Engelsen et al., 2005; EFSA NDA Panel, 2012a). The synthesis of vitamin D₃ in the skin is a function of the amount of UV-B irradiation reaching the dermis, the availability of 7-DHC and heat (body temperature). During summer months or following exposure to artificial UV-B irradiation, the synthesis of vitamin D₃ in the skin may be the main source of vitamin D. Dietary intake of vitamin D is essential in case endogenous synthesis, due to insufficient UV-B exposure, is lacking or insufficient. With increasing latitude, both the qualitative and quantitative properties of sunlight are not sufficient in parts of the year for vitamin D₃ synthesis in the skin to take place, leading to the so-called vitamin D winter (Engelsen et al., 2005). For example, in Rome, Italy (41.9°N), the vitamin D winter was reported to be from November through February; in Berlin, Germany (52.5°N) or Amsterdam, the Netherlands (52.4°N), it was reported to be between October and April (Tsiaras and Weinstock, 2011); and in Tromsø, Norway (69.4°N), it was reported to be between beginning of October through mid-March (Engelsen et al., 2005); this is based on different models and assuming a cloudless day/clear atmosphere.

Besides considering latitude and season, a UV-index can be used to estimate vitamin D₃ synthesis in the skin (Brouwer-Brolsma et al., 2016) (Section 5.3.2.1), assuming that sun exposure with a UV-index < 3 does not supply the body with sufficient vitamin D (Webb and Engelsen, 2006; McKenzie et al., 2009). The categorisation of studies where subjects are exposed to a UV-index < 3 and ≥ 3 can be done using data from the World Health Organization (WHO) (Section 5.3.2.1).4 However, it has been found that, even when the UV-index is < 3, there may be endogenous vitamin D synthesis (Seckmeyer et al., 2013). Another approach to estimate vitamin D₃ synthesis in the skin is to use a simulation model that implies a number of assumptions for the calculations (Webb, 2006; Webb and Engelsen, 2006). This simulation model estimates for example that the exposure to UV-B irradiation at 45°N at any time of the year in the middle of the day may result in vitamin D synthesis in the skin, while at 50°N, it estimates that there is no appreciable vitamin D synthesis from mid-November till February (Brouwer-Brolsma et al., 2016) (Section 5.3.2.1).

In addition to latitude and season, the vitamin D synthesis in the skin of humans is affected by several other external factors. The ozone layer effectively absorbs UV-B irradiation. Clouds, when completely overcast, can attenuate the UV-B irradiation by as much as 99%. Surface, especially snow, can however reflect up to 95% of the UV-B irradiation. Time spent outdoors, the use of sunscreen, and clothing also affect the sun-induced vitamin D synthesis in the skin (Engelsen, 2010).

After adjustment for potential confounders, individuals with initially lower serum 25(OH)D concentration (below 37.5 nmol/L) responded more quickly to UV-B exposure (and thus synthesised vitamin D in the skin) than individuals with higher concentrations (Brustad et al., 2007). The sun-induced vitamin D synthesis was reported to be higher in subjects with light skin compared to people with dark skin, because of the higher content of melanin in the skin of the latter group (Webb and Engelsen, 2006; Brouwer-Brolsma et al., 2016). The ability to vitamin D synthesis in the skin decreases with age (Lamberg-Allardt, 1984; MacLaughlin and Holick, 1985).

UV-B irradiation regulates total synthesis of vitamin D₃ in the skin, as both previtamin D₃ and vitamin D₃ present in the skin are photodegraded to biologically inert isomers following UV-B exposure (Webb et al., 1989). This downregulation of vitamin D synthesis in the skin prevents vitamin D toxicity due to prolonged sun exposure (Holick, 1994). Vitamin D intoxication by UV-B irradiation has not been reported.

The Panel notes that sun exposure may contribute a considerable and varying amount of vitamin D available to the body. The Panel considers that the association between vitamin D intake and status for the purpose of deriving dietary reference values (DRVs) for vitamin D should be assessed under conditions of minimal endogenous vitamin D synthesis (Section 5.3.2).

2.3.2 Intestinal absorption

Vitamin D from foods is absorbed throughout the small intestine, mostly in the distal small intestine. Studies using radiolabelled compounds indicate that the absorption efficiency of vitamin D varies between 55% and 99% (mean 78%) in humans, with no discrimination between vitamin D₂ and D₃ (Thompson et al., 1966; Lo et al., 1985; Jones, 2014; Borel et al., 2015; Reboul, 2015).

Due to the fat soluble characteristics of vitamin D, the absorption process is more efficient in the presence of biliary salts and when dietary fat is present in the lumen of the small intestine. A systematic review on a limited number of studies (generally reporting not statistically significant results) suggests that an oil vehicle improves the absorption of vitamin D, as shown by a greater serum 25(OH)D response, compared with a powder or an ethanol vehicle (Grossmann and Tangpricha, 2010). However, few data on the effect of the food matrix on vitamin D absorption (vitamin D₂ or vitamin D₃) have been published and the effect of the supplement matrix is not clear, as reviewed by Borel et al. (2015). A recent study reports that vitamin D₂ when given as supplement was more effective in increasing serum 25(OH)D₂ than vitamin D₂-fortified bread (Itkonen et al., 2016). Data suggest that age per se has no effect on vitamin D absorption efficiency (Borel et al., 2015). The vitamin D absorbed from the intestine is incorporated into chylomicrons that reach the systemic circulation through the lymphatic system (Jones, 2013).

The Panel considers that the average absorption of vitamin D from a usual diet is about 80%, that limited data are available on the effect of the food or supplement matrix on absorption of vitamin D (vitamin D₂ or D₃), and that age per se has no effect on vitamin D absorption efficiency.

2.3.3 Transport in blood

Transport of vitamin D from skin to storage tissue or to the liver is carried out by a specific plasma protein called vitamin D-binding protein (DBP). Transport of vitamin D₂ or D₃ from the diet to storage depots or liver is on chylomicrons, although some evidence indicates that transfer from chylomicrons to DBP occurs (Jones, 2014).

After hydroxylation of vitamin D in the liver (Section 2.3.6), serum 25(OH)D concentrations in the blood reflect the amount of vitamin D attained from both cutaneous synthesis (Section 2.3.1) and dietary sources (Section 2.3.2). In the blood, 85–90% of 25(OH)D is transported bound to DBP, 10–15% is bound to albumin, and < 1% is free (Bikle et al., 1985; Powe et al., 2013; Chun et al., 2014; Yousefzadeh et al., 2014). In a second hydroxylation step, which takes place mainly in the kidney, but also in other tissues, 1,25(OH)₂D may be formed (Section 2.3.6). In the blood, 1,25(OH)₂D is primarily transported bound to DBP and albumin (Bikle et al., 1986; Jones et al., 1998; Powe et al., 2013).

The serum concentration of 25(OH)D is approximately 1,000 times higher than that of 1,25(OH)₂D. An overview of reported 25(OH)D concentrations from studies in 17 European countries (Spiro and Buttriss, 2014) and other recent European data ((Thiering et al., 2015) in Germany) show that mean/median concentrations (Section 2.4.1) range from about 20 to 95 nmol/L in adults or children.

While serum 25(OH)D has a mean half-life of approximately 13–15 days (Jones KS et al., 2014) (Section 2.4.1) due to its strong affinity for DBP, serum 1,25(OH)₂D has a half-life measured in hours (Jones et al., 1998; IOM, 2011).

2.3.4 Distribution to tissues

Within hours of ingestion (Section 2.3.2) or synthesis in the skin (Section 2.3.1), vitamin D is distributed to the liver for conversion (hydroxylation, Section 2.3.6, Figure 2) or delivered as either vitamin D or its metabolites to the storage tissues (Section 2.3.5) (Jones, 2013). The vitamin D from dietary sources is released from the chylomicrons by action of the enzyme lipoprotein lipase upon arrival in the tissues. Serum 25(OH)D and 1,25(OH)₂D are released from DBP to various tissues such as bone, intestine, kidney, pancreas, brain and skin. Upon release from DBP, 1,25(OH)₂D is bound intracellularly to VDR (Section 2.3.6) (Gropper et al., 2009). 25(OH)D is taken up from the blood into tissues, probably by protein-binding (Mawer et al., 1972).

2.3.5 Storage

The long-term storage sites of vitamin D include mainly the adipose tissue, muscle, liver and other tissues (Heaney et al., 2009; Whiting et al., 2013).

Adipose tissue is a major repository in the body for vitamin D (Blum et al., 2008) and, in subjects with no vitamin D₂ supplementation, vitamin D was found in adipocyte lipid droplets as both vitamin D₃ and its metabolites (25(OH)D₃ and 1,25(OH)₂D₃) (Malmberg et al., 2014).

Studies have reported an inverse relationship between body mass index (BMI)/body fat and serum 25(OH)D concentrations (for reviews: Saneei et al., 2013; Vanlint, 2013). The mechanisms for this relationship are not fully understood. They have been suggested, among others, to include a 'trapping'/sequestration of vitamin D in the body tissues, particularly in adipose tissue in overweight and obese individuals (Wortsman et al., 2000; Parikh et al., 2004; Blum et al., 2008; Jungert et al., 2012), a volumetric dilution of the vitamin D in obese subjects (Drincic et al., 2012), and altered behaviour of obese subjects resulting in less cutaneous vitamin D synthesis in the skin (Vanlint, 2013).

2.3.6 Metabolism

Activation of vitamin D involves two steps. The first occurs after vitamin D is released from DBP to the liver, where it undergoes 25-hydroxylation to 25(OH)D (IOM, 2011; Jones, 2014) (Figure 2). Both a mitochondrial enzyme (CYP27A1) and several microsomal enzymes (including CYP2R1, CYP3A4 and CYP2J3) are able to carry out the 25-hydroxylation of vitamin D₂ or vitamin D₃ (Jones G et al., 2014). The 25-hydroxylation is more efficient with 'low' serum 1,25(OH)₂D concentrations than with 'normal' serum 1,25(OH)₂D concentrations (Gropper et al., 2009). The product of the 25-hydroxylation step, 25(OH)D, is bound to DBP (Section 2.3.3) and transported to the kidneys.

The second step is the 1α-hydroxylation of 25(OH)D primarily in the kidneys (Jones, 2014). Apart from the kidneys, 1,25(OH)₂D is also produced in an autocrine way in other organs such as bone cells and parathyroid cells. The placenta is one of the extrarenal sites for production of 1,25(OH)₂D by the 1α-hydroxylase. This local production supports the calcium demand of the fetus and does not contribute to the circulating concentration of 1,25(OH)₂D of the mother (Jones, 2014).

The activity of the 1α-hydroxylase (Section 2.2.1) is regulated by calcium, phosphate, and their regulating hormones (Figure 2). Any interruption of this conversion process, due to, for example, liver or kidney disease, may lead to vitamin D deficiency (Section 2.2.2.1) (Holick, 2007). After its production, 1,25(OH)₂D is transported bound to DBP in the blood (Section 2.3.3) to the target tissues (Section 2.2.1).

The metabolite 1,25(OH)₂D is fairly unstable (Section 2.3.3) without the attachment to carrier proteins (Norman, 2008; Lehmann and Meurer, 2010). Once at the target cells, 1,25(OH)₂D must be released from the DBP and current evidence suggests that it is the unbound fraction that has access to the target cells (Section 2.4). Free 1,25(OH)₂D taken up by target cells is either rapidly metabolised or bound to VDRs (Lehmann and Meurer, 2010). VDRs are involved in various regulatory processes that stand beyond classical homeostasis of calcium and phosphate. VDRs have been identified in the cardiovascular system and most cell types in the immune system, and also in other tissues like pancreas, skeletal muscle, lung, central nervous system, and reproductive system (Holick, 2004; Bischoff-Ferrari, 2010). Thus, 1,25(OH)₂D in association with VDR has a biological function not limited to bone, intestine, kidneys and parathyroid glands, but throughout the body, regulating many functions.

Upon binding of 1,25(OH)₂D, the VDR undergoes conformational changes that will allow interaction with several other transcriptional factors within the nucleus in the target cells (Bouillon et al., 2008). To interact with transcriptional factors and affect gene transcription, the active VDR must form a heterodimer with the retinoid receptor, and this heterodimer can then bind to selector or promoter sites of the target cell DNA. This new complex recruits various activators and co-depressors that affect gene expression. This can include protein synthesis and secretion, cellular proliferation or differentiation. Several factors determine the overall cellular responses, including cell type and cell number, availability of VDR and the affinity of the 1,25(OH)₂D to this receptor (Jones et al., 1998).

According to the review by Jones (2013), although vitamin D₂ and D₃ present structural differences (Figure 1, Section 2.1), qualitatively, they trigger an identical set of biological responses in the body (Figure 2), primarily by the regulation of gene expression mediated by the same VDR. None of the steps in the specific vitamin D signal transduction cascade appears to discriminate between vitamin D₂ and vitamin D₃ at the molecular level (Jones, 2013). Vitamins D₂ and D₃ are considered biologically equivalent in terms of their ability to cure rickets (Jones, 2013).

Potential differences in the biological potencies of vitamin D₂ and D₃ have been addressed in studies that measured increases in plasma 25(OH)D concentrations (Section 2.4.1) as a surrogate non-functional marker of biological activity after supplemental vitamin D₂ or D₃ (Jones, 2013; Lehmann et al., 2013; Itkonen et al., 2016). These studies have consistently shown that administration of vitamin D₂ supplements decreases the percentage contribution of vitamin D₃ to the total pool of vitamin D undergoing 25-hydroxylation, and that this decrease is accompanied by a fall in absolute serum 25(OH)D₃ concentrations. Data suggest that vitamin D₃ may be the preferred substrate for hepatic 25-hydroxylation (conversion to 25(OH)D) (Holmberg et al., 1986; Tripkovic et al., 2012), while data from toxicity and repletion studies suggest some preferential non-specific catabolism of vitamin D₂ (compared to vitamin D₃), accelerating its degradation (Jones, 2013). A meta-analysis comparing supplementation studies with vitamin D₂ and D₃ concluded that, even though bolus doses of vitamin D₃ (> 125 μg/day) were more efficacious for raising total serum 25(OH)D concentration compared with vitamin D₂ doses, the differences between the two forms of vitamin D supplements disappeared when given as lower daily doses (Tripkovic et al., 2012).

The catabolism of 25(OH)D and 1,25(OH)₂D in the body involves inactivation by 24-hydroxylation, which gives rise initially to 24,25(OH)₂D (preventing the conversion of 25(OH)D to 1,25(OH)₂D (Jones et al., 2012; Biancuzzo et al., 2013)) and to 1,24,25(OH)₃D (i.e. 1,24,25-trihydroxyvitamin D, then leading to calcitroic acid) (Section 2.3.7). Following vitamin D supplementation, 24-hydroxylase (CYP27A1) is upregulated with a lag of several weeks (Wagner et al., 2011).

There is some evidence that certain products of the degradation pathway are functional. For example, the 24,25(OH)₂D₃ is of importance in bone mineralisation and PTH suppression (Jones, 2014). Others have indicated that the 24-hydroxylated metabolites are important in fracture repair, although the vast majority of the evidence points towards 24-hydroxylation being a step in the pathway of inactivation (Jones, 2014).

The Panel notes that 1,25(OH)₂D in association with VDR has a biological function not limited to the bone, intestine, kidneys and parathyroid glands, but throughout the body, regulating many functions. The Panel also notes the conflicting results regarding the potential differences in the biological potencies and catabolism of vitamin D₂ and D₃. The Panel thus considers that the association between vitamin D intake and status for the purpose of deriving DRVs for vitamin D, may need to be investigated considering vitamin D₂ and D₃ separately (Section 5.3.2).

2.3.7 Elimination

There are two main pathways of degradation, the C23 lactone pathway, and the C24 oxidation pathway (Section 2.3.6 and Figure 2) (Holick, 1999; Jones, 2014). Vitamin D metabolites in the body are degraded in an oxidative pathway involving stepwise side-chain modifications by the actions of CYP24A1 (24-hydroxylase). 1,25(OH)₂D is a strong controller of its own degradation by stimulating the 24-hydroxylase (IOM, 2011). After several steps, one of the final products of the C24 oxidation pathway, i.e. calcitroic acid, is excreted, mainly in the bile and thus in the faeces. Human CYP24A1 also catalyses, although to a lesser extent, the 23-hydroxylation of both 25(OH)D and 1,25(OH)₂D leading, in sequential steps, to 25(OH)D-26,23-lactone and 1,25(OH)₂D-26,23-lactone, respectively (Jones G et al., 2014). 1,25(OH)₂D can also be epimerised by the conversion of the configuration of the hydroxyl-group at the C-3 of the A ring to 3-epi-1α,25(OH)₂D. Other vitamin D metabolites can be epimerised as well and are then less biologically active. 3-epi-1α,25(OH)₂D showed some transcriptional activity towards target genes and induction of antiproliferative/differentiation activity in human leukaemia cells (Kamao et al., 2004).

2.3.7.1 Faeces and urine

The majority (around 70%) of the metabolites of the vitamin D pathways of degradation are excreted in the bile (Jones, 2014). Due to active renal reuptake, the urinary excretion of vitamin D metabolites is low.

The Panel notes that the main route of excretion of vitamin D metabolites is via the faeces.

2.3.7.2 Breast milk

Breast milk accounts for a small part of the vitamin D elimination in lactating women (Taylor et al., 2013). The concentration of vitamin D in breast milk is higher than that of 25(OH)D (and of 1,25(OH)₂D), and vitamin D passes more readily from the circulation into the breast milk than 25(OH)D (Makin et al., 1983; Hollis et al., 1986). In general, mean vitamin D concentrations in breast milk of healthy lactating women, unsupplemented or supplemented with vitamin D below the UL, are low and in the range of 0.25–2.0 μg/L (Dawodu and Tsang, 2012; EFSA NDA Panel, 2013). There is a general agreement that human milk does not contain sufficient vitamin D to prevent rickets in the breast-fed infant (Olafsdottir et al., 2001).

The amount of vitamin D in human milk modestly correlates with maternal vitamin D intake up to about 18 μg/day, with evidence for a lower response in African-American compared to Caucasian women (who had mean maternal serum 25(OH)D concentration of about 67 and 112 nmol/L, respectively) (Specker et al., 1985; EFSA NDA Panel, 2012a).

Vitamin D supplementation starting in late pregnancy (i.e. after 27 weeks of gestation) (Wall et al., 2016) or early lactation (Ala-Houhala et al., 1988a; Hollis and Wagner, 2004a) may increase the vitamin D concentration of breast milk, although only modestly unless high supplemental doses are used. For example, Hollis and Wagner (2004a) supplemented 18 lactating mothers within 1 month after birth with 10 μg vitamin D₃ and with either 40 μg or 90 μg vitamin D₂ daily for 3 months. Mean serum total 25(OH)D concentration increased compared to baseline in both groups (from about 69 to about 90 nmol/L, and from about 82 to about 111 nmol/L, respectively). Mean milk antirachitic activity5 increased from 35.5 to 69.7 IU/L in the group receiving 50 μg vitamin D/day and from 40.4 to 134.6 IU/L in the group receiving 100 μg vitamin D/day. This was attributable to increases in milk concentrations of both vitamin D and 25(OH)D.

Considering a mean milk transfer of 0.8 L/day during the first 6 months of lactation in exclusively breastfeeding women (Butte et al., 2002; FAO/WHO/UNU, 2004; EFSA NDA Panel, 2009), and a concentration of vitamin D in mature human milk of 1.1 μg/L (mid-point of the range of means of 0.25–2.0 μg/L), the secretion of vitamin D into milk during lactation is around 0.9 μg/day.

The Panel considers that secretion of vitamin D into breast milk during the first 6 months of exclusive breastfeeding is about 0.9 μg/day.

2.3.8 Metabolic links with other nutrients

Vitamin D interacts with other nutrients from the diet. There is interaction between 1,25(OH)₂D, calcium and phosphorus that affects mineral and vitamin D metabolism (EFSA NDA Panel, 2015b,c). Administration of potassium salts may alter renal synthesis of 1,25(OH)₂-vitamin D (Sebastian et al., 1990; Lemann et al., 1991). Vitamin A has been suggested to interfere with the action of vitamin D. The active metabolite of vitamin A, i.e. retinoic acid, and 1,25(OH)₂D regulate gene expression through nuclear receptors (Section 2.3.6). Data on interactions between vitamin A and vitamin D have been reviewed (SCF, 2002b; EFSA NDA Panel, 2015a). Both 1,25(OH)₂D and vitamin K are needed for the synthesis of osteocalcin in the osteoblasts and 1,25(OH)₂D regulates the expression of osteocalcin.

2.4 Biomarkers

2.4.1 Plasma/serum concentration of 25(OH)D

Plasma or serum concentration of 25(OH)D represents total vitamin D from exposure to both UV irradiation (cutaneous synthesis) and dietary sources (Section 2.3.3) and can be used as a biomarker of vitamin D intake in people with low exposure to UV-B irradiation from sunlight (EFSA NDA Panel, 2012a). Serum 25(OH)D has a long mean half-life of approximately 13–15 days (IOM, 2011; Jones KS et al., 2014) (Section 2.3.3) and is considered a useful marker of vitamin D status (both D₂ and D₃) (Seamans and Cashman, 2009; EFSA NDA Panel, 2012a).

Plasma/serum 25(OH)D₂ is of dietary origin only, while plasma/serum 25(OH)D₃ may be of dietary or dermal origin (Sections 2.3.1 and 3). Body composition has an impact on serum 25(OH)D concentration and an inverse correlation between serum 25(OH)D concentrations and BMI has been observed (Section 2.3.5). Increasing oral vitamin D intake increases 25(OH)D concentration until a plateau is reached after about 6 weeks, which indicates an equilibrium between the production and degradation of serum 25(OH)D (Vieth, 1999; Viljakainen et al., 2006b).

A linear relationship was reported between vitamin D intake and serum 25(OH)D concentrations up to a total vitamin D intake of 35 μg/day (Cashman et al., 2011b) and 50 μg/day (Cranney et al., 2007). The US Institute of Medicine (IOM, 2011) found a steeper rise in the serum 25(OH)D concentrations with vitamin D intakes up to 25 μg/day and a slower, more flattened response when 25 μg/day or more were consumed (Section 5.3.2).

During pregnancy, maternal 25(OH)D concentration is generally unaffected, according to most of (but not all) the studies reviewed by IOM (2011) for this aspect (Section 4, Appendix B). The contradictory results in the literature are also discussed by others (Zhang et al., 2014). In this longitudinal analysis of a random sample (not seasonally balanced) of 30 women between 15 weeks of gestation and 2 months post partum, mean serum total 25(OH)D concentration significantly decreased between 20 and 36 weeks of gestation (from 51.0 to 37.4 nmol/L).

There is an ongoing debate about the optimal range of serum 25(OH)D concentration and the cut-off values for defining deficiency, insufficiency and sufficiency (Jones, 2014) (Section 4). A serum 25(OH)D concentration of 25–30 nmol/L has been proposed as a value below which the risk of rickets and osteomalacia increases (Cashman et al., 2011a). Other health outcomes may also be considered (Sections 4 and 5.1).

There are numerous methods for the measurement of 25(OH)D in serum (Wallace et al., 2010; Carter, 2011) including high-performance liquid chromatography with UV-detection (HPLC/UV), liquid chromatography-tandem mass spectrometry (LC–MS/MS), and immunoassays (radioimmunoassays RIA, competitive protein binding assays CPBA, enzyme-linked immunosorbent assays ELISA) that are either manual or automated. LC–MS/MS and HPLC methods are considered the gold standard methods (Wallace et al., 2010; Carter, 2011). These methods have the advantage that they can measure 25(OH)D₃ and 25(OH)D₂ separately, which is needed in specific situations (Tai et al., 2010; Carter, 2011). Also, some methods allow detection of other vitamin D metabolites, such as 24,25(OH)₂D (Wallace et al., 2010; Carter, 2012).

Formerly, all methods suffered from the lack of an international common standard, this lack contributing to the variability of results of serum 25(OH)D measurements (Section 5.1.2.1.7). The Vitamin D External Quality Assessment Scheme (DEQAS)6 has revealed considerable differences between methods (both within and between laboratories), raising concerns about the comparability and accuracy of different assays and laboratories (Snellman et al., 2010; Carter, 2011; Farrell et al., 2012; Heijboer et al., 2012). The introduction of a standard reference material for vitamin D in human serum by the US National Institute of Standards and Technology (NIST)7 has been a step forward in providing a reference measurement procedure (RMP) against which assays could be standardised (Carter, 2012). The Vitamin D Standardization Program (VDSP)8 has developed protocols for standardising procedures of 25(OH)D measurement in national health/nutrition surveys to promote 25(OH)D measurements that are accurate and comparable over time, location, and laboratory to improve public health practice (Cashman et al., 2013). The VDSP RMP has been joined by a number of commercial methods and laboratories and thus, their results are comparable to LC–MS/MS as regards 25(OH)D concentrations. In the VDSP, LC–MS/MS is the reference method. According to a reanalysis of serum 25(OH)D concentrations using the VDSP protocol, the range of mean concentrations (Section 2.3.3) in 14 European studies in children and adult populations (including one study in migrants in Finland) was 38.3–65 nmol/L (versus 44.8–69 nmol/L in the originally analysed serum 25(OH)D data) (Cashman et al., 2016).

Thus, there is a range of methodologies available for the measurement of 25(OH)D, and each method has its advantages and limitations (Wallace et al., 2010). Given the lack of consensus on the optimal range of serum 25(OH)D concentration and on the cut-off values for defining deficiency, insufficiency and sufficiency mentioned above, the Panel considered relevant studies on the relationship between serum 25(OH)D concentration and health outcomes (Section 5.1), and this review was undertaken irrespective of the analytical method applied to measure serum 25(OH)D concentration. However, analytical methods are considered by the Panel in a sensitivity analysis for the assessment of the relationship between total vitamin D intake and serum 25(OH)D concentration (Section 5.3.2, Appendices C and D).

2.4.2 Free serum 25(OH)D concentration

Free serum 25(OH)D is the fraction of serum 25(OH)D that circulates without being bound to DBP and albumin (Section 2.3.3). This free form accounts for less than 1% of total 25(OH)D in the body, but has been hypothesised to be a potential marker of vitamin D status, because this free fraction is readily available to target cells (Powe et al., 2013; Chun et al., 2014; Johnsen et al., 2014).

The Panel considers that, at present, free serum 25(OH)D concentration cannot be used as a biomarker of vitamin D intake and status and that more research is needed to establish the potential of free serum 25(OH)D concentration as a biomarker of vitamin D status.

2.4.3 Plasma/serum 1,25(OH)₂D concentration

The biologically active 1,25(OH)₂D has a half-life measured in hours (Section 2.3.3) and is closely linked with blood calcium, PTH, and phosphate concentrations (Sections 2.2.1 and 2.3.6, Figure 2). During pregnancy, maternal 1,25(OH)₂D synthesis increases (IOM, 2011) (Section 2.4.1).

Zerwekh (2008) considered that plasma/serum 1,25(OH)₂D concentration cannot be used to assess vitamin D status, in view of its short half-life and the tight regulation of its concentration. Serum 1,25(OH)₂D concentrations do not change according to month of the year (apart in October compared to April) within serum 25(OH)D₃ concentrations of 40 nmol/L and 78 nmol/L in healthy children and adults (18 months–35 years) (Chesney et al., 1981). In a cross-sectional study of post-menopausal women, serum 1,25(OH)₂D concentration was found to be negatively correlated with serum 25(OH)D concentration at 25(OH)D concentrations ≤ 40 nmol/L and positively at concentrations > 40 nmol/L, illustrating a non-linear association between concentrations of serum 25(OH)D and of the active metabolite 1,25(OH)₂D (Need et al., 2000). In this study, at serum 25(OH)D concentrations ≤ 40 nmol/L (compared to higher concentrations), 1,25(OH)₂D concentration was found to be closely related to PTH concentration.

In another study of vitamin D metabolites and calcium absorption in older patients with 25(OH)D concentration < 40 nmol/L (Need et al., 2008) (Section 5.1.2.1.6 and Appendix B), serum 1,25(OH)₂D concentrations were significantly decreased concurrent with increases in serum PTH, ALP, and urine hydroxyproline in subjects with serum 25(OH)D < 10 nmol/L. This suggests that this level of substrate is insufficient to maintain serum 1,25(OH)₂D concentration, despite secondary hyperparathyroidism.

The Panel considers that, because of the tight homeostatic regulation of 1,25(OH)₂D concentration in blood, this marker cannot be used as a biomarker of vitamin D status, but rather reflects vitamin D function.

2.4.4 Serum parathyroid hormone (PTH) concentration

Serum PTH concentration and its relationship with 25(OH)D concentration (via its relationship with 1,25(OH)₂D, Sections 2.2.1, 2.3.6 and 2.4.3, and Figure 2) has been suggested as a possible biomarker or functional endpoint of vitamin D status. Sai et al. (2011) reviewed 70 studies undertaken in children or adults and showed that it was not possible to set a cut-off value for 25(OH)D concentration using PTH as a reference, due to the low consistency in the cut-off value observed in these studies. A systematic review and meta-analysis of 36 RCTs and four before–after studies that investigated vitamin D supplementation in healthy subjects and the response of 25(OH)D, PTH, BMD, bone markers and calcium absorption, revealed large heterogeneity across the results when comparing 18 RCTs using PTH as a biomarker of vitamin D status (Seamans and Cashman, 2009). In this publication, subgrouping by addition of calcium supplementation or no calcium supplementation suggested an effect of vitamin D supplementation on circulating PTH in the absence of calcium, without important heterogeneity, but not in the presence of calcium supplementation, with strong heterogeneity.

The Panel considers that serum PTH concentration is not a biomarker of vitamin D intake, as serum PTH is also influenced by, e.g. serum calcium and phosphate concentrations and other factors. The Panel also considers that serum PTH concentration in healthy subjects is not a useful biomarker for vitamin D status as assessed by serum 25(OH)D concentration.

2.4.5 Other biomarkers

Since vitamin D is a well-established nutrient in relation to bone, markers of bone formation and turnover (osteocalcin, bone specific ALP and urine N-telopeptide crosslinks) have been considered as markers of long-term status of vitamin D (Bonjour et al., 2014). Low urinary calcium excretion and an increased bone specific ALP activity have been used as biomarkers in the diagnosis of vitamin D deficiency (Section 2.2.2.1).

Serum concentrations of calcium and inorganic phosphorus that may be low and high PTH serum concentration can help in the diagnosis of rickets or osteomalacia (Section 2.2.2.1). Structural bone markers (low BMD, rickets or osteoporosis) have also been used as biomarkers of vitamin D status, but have the disadvantage of a slow reaction time, which means that when the condition is diagnosed, bone health may be irreversibly damaged.

The Panel considers that more research is needed to establish the relationship between responses of bone markers (e.g. osteocalcin, bone ALP and urine N-telopeptide crosslinks) to changes in vitamin D status.

2.4.6 Conclusions on biomarkers

The Panel considers that serum 25(OH)D concentration can be used as a biomarker of vitamin D intake in a population with low exposure to UV-B irradiation (from sunlight, Section 2.3.1), and of vitamin D status at population level. The Panel notes that, due to the high variability in 25(OH)D measurements obtained with different analytical methods (Section 2.4.1), comparison of results from different studies as well as to reference range values has to be done with caution.

2.5 Effects of genotypes

Some polymorphisms of genes encoding proteins involved in vitamin D synthesis, transport and metabolism influence serum 25(OH)D concentration (Berry and Hyppönen, 2011). Two genome-wide association studies (Ahn et al., 2010; Wang et al., 2010), conducted as meta-analyses of data from subjects of European ancestry, identified variants in the genes DHCR7, CYP2R1, GC (group specific component gene) and CYP24A1, expressing 7-dehydrocholesterol reductase (DHCR7), 25-hydroxylase, DBP and 24-hydroxylase, respectively.

Mutations in DHCR7, going along with an impaired activity of the gene, are seen in the rare Smith–Lemli–Opitz syndrome and result in an accumulation of 7-DHC (Figure 1, Sections 2.1 and 2.3.1), the substrate for the 25(OH)D synthesis in the skin (Berry and Hyppönen, 2011). It has been reported that DHCR7 mutations are related to a higher vitamin D status and that allele frequencies of DHCR7 single nucleotide polymorphisms (SNPs) are high at Northern latitudes (0.72 in Europe, 0.41 in Northeast Asia) (Kuan et al., 2013). CYP2R1 encodes the enzyme primarily responsible for the hydroxylation of vitamin D to 25(OH)D in the liver (Section 2.3.6) and GC encodes the DBP that is the major carrier protein for vitamin D and its metabolites (Section 2.3.3). Variants in both genes have been associated with lower 25(OH)D serum concentrations in carriers as compared to non-carriers (Nissen et al., 2014). However, genetic variations in the GC gene were also associated with enhanced albumin-bound and free, and therefore readily bioavailable, serum 25(OH)D concentrations (Sections 2.3.3 and 2.4.2) (Powe et al., 2013; Chun et al., 2014; Johnsen et al., 2014). Season, dietary and supplemental intake may modify the effects on serum 25(OH)D concentration of the variants in the genes GC and CYP2R1 (Engelman et al., 2013; Waterhouse et al., 2014).

CYP24A1 catalyses the conversion of both 25(OH)D₃ and 1,25(OH)₂D₃ into 24-hydroxylated products to be excreted (Sections 2.3.6 and 2.3.7). The reaction is important in the regulation of the concentration of the active 1,25(OH)₂D in the kidney and in other tissues (Jones et al., 2012). Inactivating mutations in the gene encoding this enzyme can cause idiopathic infantile hypercalcaemia (Dinour et al., 2013) and have been linked to other hypercalcaemic conditions causing nephrolithiasis and nephrocalcinosis (Jones et al., 2012). The possibility that increased expression of CYP24A1 may be an underlying cause of vitamin D deficiency and progression of disease states has been discussed (Jones et al., 2012). Associations of the CYP27B1 genotypes, that code for 1α-hydroxylase (Sections 2.2.1 and 2.3.6), with 25(OH)D concentrations have also been reported (Hypponen et al., 2009; Signorello et al., 2011) but were not found significant in other studies (Berry and Hyppönen, 2011). With regard to variants of the gene encoding VDR, there is no consistent finding on their relationship with serum 25(OH)D concentrations; in particular in some studies investigating the Fok-1 polymorphism of VDR, it is not clear how this SNP influences 25(OH)D concentrations (McGrath et al., 2010; Nieves et al., 2012).

The Panel considers that data on the effect of genotype on vitamin D synthesis, transport and metabolism are insufficient for use in deriving the requirements for vitamin D according to genotype variants.

3 Dietary sources and intake data

Vitamin D₂ and D₃ are fat-soluble. The major food sources for naturally occurring vitamin D₃ include animal foods such as fatty fish, offal (particularly liver), meat and meat products, and egg yolks (Anses/CIQUAL, 2012; Schmid and Walther, 2013).

Fish (and especially fatty fish and fish liver) have the highest natural content of vitamin D (Schmid and Walther, 2013), presumably derived from an accumulation in the food chain originating from microalgae that contain both vitamin D₃ and provitamin D₃ (Jäpelt and Jakobsen, 2013). Egg yolk also has a high vitamin D₃ content (Schmid and Walther, 2013), which strongly correlates with the content of vitamin D₃ of the hen's feed (Mattila et al., 1993, 1999). Animal studies showed that vitamin D₃ and 25(OH)D₃ were effectively transferred from the hen to the egg yolk, depending on the hen's diet (Mattila et al., 2011) and UV-B exposure (Kuhn et al., 2015). The content of vitamin D of meat products varies and depends, among other things, on the contents of vitamin D in the fodder, the fat content of the meat product, and latitude where the animals have grazed (Mattila et al., 2011; Liu et al., 2013).

The vitamin D metabolite 25(OH)D is present in some foods of animal origin in varying amounts (Mattila et al., 1993, 1995, 1999; Clausen et al., 2003; Ovesen et al., 2003; Jakobsen and Saxholt, 2009; Cashman, 2012). Due to the suggested higher biological activity of 25(OH)D in foods compared with the native vitamin D, a conversion factor of 5 has been used for 25(OH)D₃ in the calculation of total vitamin D₃ in some food composition tables, including those in the UK, Denmark and Switzerland (Cashman, 2012; Cashman et al., 2012).

Some higher fungi, such as mushrooms, are a natural source of vitamin D₂. Vitamin D₂ is produced in fungi and yeasts by UV-B exposure of provitamin D₂ and the content depends on the amount of UV-B light exposure and time of exposure (Kristensen et al., 2012; Tangpricha, 2012).

Further sources of dietary vitamin D are fortified foods (most often milk, margarine and/or butter, and breakfast cereals) and dietary supplements. Currently, cholecalciferol (vitamin D₃) and ergocalciferol (vitamin D₂) may be added to both foods9 and food supplements.10 The vitamin D content of infant and follow-on formulae and of processed cereal-based foods and baby foods for infants and children is regulated.11

The stability of vitamin D₃ and 25(OH)D₃ and vitamin D₂ in foodstuffs during cooking has been shown to vary widely with heating process and foodstuffs, with reported retentions in eggs, margarine and bread after boiling, frying and baking of between 40% and 88% (Jakobsen and Knuthsen, 2014).

Published dietary intake data (mean/median and high percentiles) have been collected for adults in 14 European countries and for infants and children in 11 European countries (EFSA NDA Panel, 2012a). Mean intakes of vitamin D in European countries varied according to sex, age and supplementation habits. A direct comparison between countries was difficult as there was a large diversity in the methodology used for dietary assessment, age classification was not uniform, and data from food composition tables used for nutrient intake estimation were different. In the data collected from the different surveys/studies considered, mean/median intake of vitamin D from foods varied from 1.1 to 8.2 μg/day in adults. It varied from 1.7 to 5.6 μg/day in children aged about 1–5 years old, from 1.4 to 2.7 μg/day in children aged about 4–13 years old, and from 1.6 to 4.0 μg/day in children aged about 11–18 years old. When foods and supplements were considered together, mean vitamin D intake varied from 3.1 to 23.5 μg/day in adults. It varied from 8.9 to 12.5 μg/day in infants, from 2.3 μg/day to 9.0 μg/day in children aged about 1.5–3 years old, and from 1.8 μg/day to 6.6 μg/day in children aged about 4–11 years old. In high consumers (95th percentile) in adults, intake was up to 16 μg/day from foods and up to about 24 μg/day from foods and supplements. In high consumers (90th or 95th percentile according to surveys) in infants, children and adolescents, intake from foods and supplements was, respectively, up to 19 μg/day, 15 μg/day and 8 μg/day (EFSA NDA Panel, 2012a).

4 Overview of dietary reference values and recommendations

4.1 Adults

The German-speaking countries (D-A-CH, 2015) considered a review (Linseisen et al., 2011) following the guidelines of the German Nutrition Society on evidence-based nutrition. A serum 25(OH)D concentration of at least 50 nmol/L was considered advisable for bone health in younger adults (aged less than 65 years), as well as in older adults (65 years and over) (Dawson-Hughes et al., 2005; Linseisen et al., 2011). For younger adults, D-A-CH reported on IOM (2011) and an Irish study undertaken in winter at latitudes comparable with those of Germany (Cashman et al., 2008), that showed that 10 or 20 μg/day of supplemental vitamin D allowed, respectively, 50% or 90–95% of the population to reach a serum 25(OH)D concentration above 50 nmol/L. For older adults, the main focus was the minimisation of the age-related loss of bone mass, the risk of bone fractures, skeletal muscle function and the related risks of loss of strength/mobility/balance, of falls and of fractures (Pfeifer et al., 2000, 2009; Bischoff-Ferrari et al., 2003; Dawson-Hughes et al., 2010; EFSA NDA Panel, 2011; IOM, 2011; Linseisen et al., 2011). D-A-CH considered that studies in older adults supported a protective effect of 10–20 μg/day supplemental vitamin D on loss of the ability to move, on falls, fractures and premature death (Autier and Gandini, 2007; Bischoff-Ferrari et al., 2009a,b; LaCroix et al., 2009; Bjelakovic et al., 2011; Linseisen et al., 2011). With 50 μg/day vitamin D, about 90–95% of older adults had a serum 25(OH)D concentration above 50 nmol/L and 50% had a concentration of 75 nmol/L (Cashman et al., 2009). D-A-CH set the adequate intake (AI) for all adults at 20 μg/day in situations in which endogenous vitamin D synthesis is absent. D-A-CH considered vitamin D supplements and/or endogenous synthesis to cover the difference between the 'usual' intake (2–4 μg/day) and this value.

The Nordic Council of Ministers (2014)12 considered a systematic literature review on vitamin D intake/status and health (Lamberg-Allardt et al., 2013) (Section 5.1), based on which a serum 25(OH)D concentration of 50 nmol/L was considered as indicative of a sufficient vitamin D status in adults. They also reported on a systematic review of intervention studies on vitamin D supplementation (Cashman et al., 2011a), from which five studies (Ala-Houhala et al., 1988b; Barnes et al., 2006; Cashman et al., 2008, 2011b; Viljakainen et al., 2009) were used for specific meta-regression analyses (Section 5.3.1). Based on two meta-regression analyses in different age groups (Section 5.3.1), the average requirement (AR) for all adults and the Recommended Intake (RI) for adults aged less than 75 years were set at 7.5 and 10 μg/day, respectively, assuming some contribution of endogenous synthesis of vitamin D during outdoor activities in summer. An RI was set at 20 μg/day for people with little or no sun exposure during the summer as well as for adults aged 75 years and over, to account for their more limited endogenous synthesis and in consideration of the available data on total mortality, bone health, fractures and falls. A lower intake level of 2.5 μg/day was also set.

The Health Council of the Netherlands (2012) considered that diet provides one-third of the vitamin D requirement and sufficient sun exposure provides the remainder. It was roughly estimated that sunlight exposure in the Netherlands leads to the production, on average over the whole year, of 6–7 μg/day vitamin D, with significantly higher amounts during summer and a decline to nearly zero during winter (Signaleringscommissie Kanker van KWF Kankerbestrijding, 2010). The Council considered that an intake of 11–15 μg/day (Cashman et al., 2008, 2009, 2011b) would be sufficient for adults aged < 70 years to reach a serum 25(OH)D concentration > 30 nmol/L that was derived from data on prevention of rickets in young children. As there was no sign that vitamin D supplementation is required in this group, the Council rounded the AI down to 10 μg/day. Adults with fair skin and insufficient sun exposure, or with dark skin, or women aged 50–70 years regardless of skin colour and amount of time spent outdoors, were advised to take a vitamin D supplement of 10 μg/day. In older adults (≥ 70 years), an intake of 20–25 μg/day was considered sufficient to reach a serum 25(OH)D concentration of 50 nmol/L, which was considered advisable for protection against bone fractures (Health Council of the Netherlands, 2000; Cranney et al., 2007; Chung et al., 2009; IOM, 2011). Considering age-related physiological changes (IOM, 2011), for older adults (70 years and over), an estimated average requirement (EAR) and a recommended dietary allowance (RDA) of 10 and 20 μg/day were set. As sun exposure and dietary intake of vitamin D vary in this age group, all older adults were advised to take a vitamin D supplement of 20 μg/day. All these reference values apply in case of insufficient exposure to sunlight.

IOM (2011) (Appendix B) underlined the interactions between calcium and vitamin D with regard to bone health and the lack of a dose–response relationship between vitamin D intake and bone health. However, based on systematic reviews (Cranney et al., 2007; Chung et al., 2009) and other data published after them, IOM considered that total vitamin D intake can be related to change in serum 25(OH)D concentrations under minimal sun exposure and that a dose–response curve for serum 25(OH)D and bone health outcomes can be established. It was considered that serum 25(OH)D concentrations below 30 nmol/L were associated with an increased risk of rickets, impaired fractional calcium absorption and decreased bone mineral content (BMC), in children and adolescents. Concentrations below 30 nmol/L were also associated with an increased risk of osteomalacia and impaired fetal skeletal outcomes, impaired fractional calcium absorption and increased risk of osteomalacia in young and middle-aged adults, and impaired fractional calcium absorption and fracture risk in older adults (IOM, 2011). The IOM considered serum 25(OH)D concentrations > 50 nmol/L as adequate for good bone health for practically all individuals. From the dose–response curve for serum 25(OH)D and bone health outcomes, assuming a normal distribution of requirements, the IOM selected serum 25(OH)D concentrations of 50 nmol/L, 40 nmol/L and 30 nmol/L as, respectively, the 'RDA-type' and 'EAR-type' reference values, and the 'lower end of the requirement range'. The IOM undertook specific meta-regression analyses (Section 5.3.1). From the lack of effect of age in these analyses, the IOM concluded that the intake to achieve the EAR-type value of 40 nmol/L was the same across all populations considered. From these analyses, an intake of 10 and 15 μg/day vitamin D would predict a mean serum 25(OH)D concentration higher than the EAR and RDA-type values in children and adults, but given the uncertainties of the analyses, these intakes were selected for the EAR (all adults) and the RDA (until the age of 70 years). For ages 51–70 years, the IOM found no basis to set a specific RDA, as women of this age may have some degree of bone loss but a lower fracture risk than later in life, and as there was generally no effect of vitamin D alone on bone health in this age group. Given the diversity of physiological/health status of adults older than 70 years, and uncertainties and variabilities in the physiology of ageing, IOM set the RDA at 20 μg/day, considering the reported significant effect of 2.5 mg of vitamin D every 4 months (equivalent to 20 μg/day) on the relative risk of fracture in (mainly) men (without calcium supplementation) (Trivedi et al., 2003).

WHO/FAO (2004) considered that a serum 25(OH)D concentration above 27 nmol/L ensures normal bone health. WHO/FAO (2004) reported on the previous approach of IOM (1997) and calculated the recommended nutrient intakes by doubling the vitamin D dietary intake (rounded to the nearest 1.25 μg) required to maintain 25(OH)D concentrations above 27 nmol/L, in order to cover the needs of all individuals irrespective of sunlight exposure. Between 42°N and 42°S, the most efficient way to acquire vitamin D was considered to usually be the endogenous synthesis in the skin. About 30 min of daily sun exposure of the arms and face without sunscreen could usually provide the daily vitamin D needs (Holick, 1994). Subjects not synthesising vitamin D because of factors such as latitude, season (particularly winter at latitudes higher than 42°), ageing, skin pigmentation, clothing, or sunscreen use, were recommended to consume the recommended nutrient intake. WHO/FAO mentioned the age-related decline in the rate of vitamin D synthesis in the skin, in the rate of vitamin D hydroxylation and in the response of target tissues such as bone (Holick, 1994; Shearer, 1997). WHO/FAO also mentioned studies in older adults, including institutionalised subjects or inpatients with low sun exposure, reporting on 'low' 25(OH)D and elevated PTH or ALP concentrations in plasma, decline in bone mass and increase in the incidence of hip fractures (Chapuy and Meunier, 1997; Dawson-Hughes et al., 1997). The recommended nutrient intakes for adults were set at 5 μg/day (19–50 years), 10 μg/day (51–65 years) and 15 μg/day (> 65 years).

The French food safety agency (Afssa, 2001) estimated vitamin D requirements to be 10–15 μg/day from the minimal amounts needed to prevent or correct deficiency (Holick, 1994, 1998; Glerup et al., 2000), and estimated endogenous production to cover 50–70% of these requirements in case of 'normal' sun exposure (i.e. about 5–7 μg/day), thus the reference value was set at 5 μg/day for adults aged < 75 years. For older adults, sun exposure was reported to be frequently insufficient (particularly in women in summer), intestinal absorption to be reduced and endogenous production to be less efficient (Dawson-Hughes, 1996). Considering seasonal changes in serum 25(OH)D and serum PTH concentrations and bone health in older adults (Dawson-Hughes, 1996; Cynober et al., 2000), the reference value was set at 10–15 μg/day. This was higher than the spontaneous intake observed at that time in France (ESVITAF, 1986; Hercberg et al., 1994), therefore the consumption of supplements under medical supervision or of fortified foods was discussed. The importance of calcium intake was also stressed.

SCF (1993) considered serum 25(OH)D concentration ranges of 25–100 nmol/L (whole population) and 25–50 nmol/L (older and institutionalised people) as advisable. The dietary vitamin D intake needed to attain serum 25(OH)D concentration of 25–100 nmol/L was considered to depend on, e.g. latitude, climate, air pollution, social and ethnic groups in Europe, and considered this intake not to be essential for healthy adults with appropriate calcium and phosphate intake and sun exposure (Markestad and Elzouki, 1991). The SCF lacked data on the effect of dietary vitamin D on serum 25(OH)D concentration of non-pregnant young adults. Based on studies on older adults (65 years and over) (MacLennan and Hamilton, 1977; Toss et al., 1983), an intake of 10 μg/day was considered to maintain 25(OH)D concentrations of 25–100 nmol/L, even in case of minimal endogenous synthesis. For adults aged 18–64 years, the acceptable range of intake was 0–10 μg/day (the highest value being set in case of minimal endogenous vitamin D synthesis). Because of lack of sun exposure and the decline with age of endogenous vitamin D synthesis, the SCF considered older adults (65 years and over) and institutionalised people to require 10 μg/day of vitamin D to maintain 25(OH)D concentrations of 25–50 nmol/L (MacLennan and Hamilton, 1977; Toss et al., 1983).

The UK has revised the DRVs for vitamin D (DH, 1991), reviewing available data on vitamin D intake/serum 25(OH)D concentration and musculoskeletal health outcomes, i.e. rickets in infants and children (Section 4.2); osteomalacia in adults; 'bone health indices' (BMD, BMC, biochemical markers of bone turnover), muscle strength and function for different life stages; stress fractures in younger adults (< 50 years); fracture prevention and risk of falls in adults aged more than 50 years; as well as non-musculoskeletal health outcomes (SACN, 2016). Based on this review, the Scientific Advisory Committee on Nutrition (SACN) considered that the risk of poor musculoskeletal health is increased at serum 25(OH)D concentrations below about 20–30 nmol/L and retained the concentration of 25 nmol/L as set by DH (1998). This concentration of 25 nmol/L was considered to 'represent a 'population protective level'; i.e. the concentration that individuals in the UK should be above, throughout the year, in terms of protecting musculoskeletal health'. A Reference Nutrient Intake (RNI) of 10 μg/day,13 applicable throughout the year, was set for the UK population aged 4 years and over (Cashman et al., 2008, 2009, 2011b). This was considered as 'the average amount needed to achieve a serum 25(OH)D concentration ≥ 25 nmol/L during winter in 97.5% of the population'. It also applies to 'population groups at increased risk of having a serum 25(OH)D concentration < 25 nmol/L', e.g. minority ethnic groups with darker skin.

An overview of DRVs for vitamin D for adults is presented in Table 1.

Table 1. Overview of dietary reference values for vitamin D for adults

	SACN (2016)	D-A-CH (2015)	NCM (2014)	NL (2012)i ⁱ NL, Health Council of the Netherlands	IOM (2011)	WHO/FAO (2004)a ^a PRI in case of no endogenous vitamin D synthesis.	Afssa (2001)	SCF (1993)	DH (1991)b ^b DRVs revised by SACN (2016).
Age (years)	≥ 18	≥ 19	18–74	18–69	19–70	19–50	20–74	18–64	19–64
DRV (μg/day)	10c ^c PRI.	20d ^d AI in case of lack of endogenous synthesis.	10e ^e PRI assuming some endogenous vitamin D synthesis. PRI of 20 μg/day in case of little or no sun exposure during the summer season.	10d ^d AI in case of lack of endogenous synthesis.	15f ^f PRI considering minimal sun exposure.	5	5g ^g Populations with 'normal' sun exposure.	0–10h ^h Acceptable range of intake. Zero in case of adequate endogenous synthesis, 10 μg/day for younger adults in case of minimal endogenous synthesis.	0
Age (years)						51–65
DRV (μg/day)						10
Age (years)			≥ 75	≥ 70	≥ 71	≥ 66	≥ 75	≥ 65	≥ 65
DRV (μg/day)			20c ^c PRI.	20c ^c PRI.	20f ^f PRI considering minimal sun exposure.	15	10–15	10	10

^a PRI in case of no endogenous vitamin D synthesis.
^b DRVs revised by SACN (2016).
^c PRI.
^d AI in case of lack of endogenous synthesis.
^e PRI assuming some endogenous vitamin D synthesis. PRI of 20 μg/day in case of little or no sun exposure during the summer season.
^f PRI considering minimal sun exposure.
^g Populations with 'normal' sun exposure.
^h Acceptable range of intake. Zero in case of adequate endogenous synthesis, 10 μg/day for younger adults in case of minimal endogenous synthesis.
ⁱ NL, Health Council of the Netherlands

4.2 Infants and children

D-A-CH (2015) considered that infants reach a serum 25(OH)D concentration of at least 50 nmol/L with an intake of 10 μg/day (Wagner et al., 2006, 2010), which was set as the AI, achieved through supplementation, independent of vitamin D endogenous synthesis and intake through consumption of breast milk or formulas. For older children, a serum 25(OH)D concentration of at least 50 nmol/L was considered to be achieved with an intake of 5–10 μg/day (Viljakainen et al., 2006b). However, a higher value of 20 μg/day was set as the AI for all children after 1 year given the lack of sun exposure (Cashman et al., 2011b) and vitamin D supplementation was recommended in winter time for children aged up to 2 years (Wabitsch et al., 2011).

The Nordic Council of Ministers (2014) set a RI of 10 μg/day up to the age of 2 years, based on rickets prevention (Markestad, 1983; Ala-Houhala, 1985; Specker et al., 1992) and the low sun exposure in Nordic countries. For older children, the vitamin D intake required for serum 25(OH)D concentration above 50 nmol/L in Danish adolescent girls throughout winter was shown to be partly dependent on the status in early autumn (Andersen et al., 2013). A meta-regression analysis on data on children and young adults (Section 5.3.1) was used to set the RI at 10 μg/day, assuming some vitamin D endogenous synthesis during summer outdoor activities.

The Health Council of the Netherlands (2012) used data on the effect of 7.5–10 μg/day supplemental vitamin D for rickets prevention (Lerch and Meissner, 2007) and assumed a sufficient calcium intake to set an AI of 10 μg/day for children aged up to 4 years. As most young children do not consume sufficient vitamin D and they should be protected against the sun, the Council advised all young children to take a 10 μg/day vitamin D supplement. Above 4 years, an AI of 10 μg/day was also set, and fair-skinned children sufficiently exposed to sunlight and with a varied diet (including low-fat margarine, cooking fats and oils) were not considered to require supplemental vitamin D.

IOM (2011) (Appendix B) considered that data were insufficient to establish an EAR for infants and that the low breast milk vitamin D concentration could not be used to set requirements. In infants, an intake of 10 μg/day was associated with no clinical deficiency and a serum 25(OH)D concentration generally above 50 nmol/L (Greer et al., 1982; Rothberg et al., 1982; Ala-Houhala, 1985; Ala-Houhala et al., 1988b; Greer and Marshall, 1989; Hollis and Wagner, 2004a). Thus, 10 μg/day was chosen as the AI for the first year of life, assuming an early supplementation of breast-fed infants and a gradual increase in formula intake in the other infants. For the age 1–18 years, IOM assumed a normal distribution of requirements and minimal sun exposure to set the same EAR and RDA as for adults aged less than 70 years (i.e. 10 and 15 μg/day, respectively).

WHO/FAO (2004) considered infants to be at risk for vitamin D deficiency because of their high skeletal growth, particularly breast-fed infants because of the low vitamin D concentration in breast milk (Specker et al., 1985) and low sun exposure. Sporadic cases of rickets in Northern cities, almost always in breast-fed infants (Binet and Kooh, 1996; Brunvand and Nordshus, 1996; Gessner et al., 1997; Pettifor and Daniels, 1997), and the increased need for 1,25(OH)₂D at puberty (Aksnes and Aarskog, 1982) were mentioned. Adolescents were considered to usually have sufficient sun exposure to synthesise vitamin D, and vitamin D produced in summer and early autumn to be stored mainly in adipose tissue (Mawer et al., 1972), thus available for winter time. However, 'low' vitamin D stores during adolescence may occur (Gultekin et al., 1987). WHO/FAO set a recommended nutrient intake of 5 μg/day for infants and children with insufficient vitamin D synthesis (e.g. during winter at latitudes higher than 42°).

Afssa (2001) set the reference value at 20–25 μg/day for infants, taking into account the frequency of rickets in some French regions and of 'low' 25(OH)D concentrations at the end of winter. The reference values for children were set at 10 μg/day (1–3 years), and then at 5 μg/day (4–19 years) based on the same considerations as for adults. Supplementation of breast-fed and formula-fed infants (10–20 μg/day), of children aged 18 months to 5 years during winter (10–20 μg/day), and of adolescents during winter and with low sun exposure (Zeghoud et al., 1995) was advised.

SCF (1993) considered the incidence of rickets in unsupplemented infants and serum 25(OH)D concentrations in supplemented and unsupplemented infants (Poskitt et al., 1979; Garabedian et al., 1991). The SCF considered that infants 6–11 months should consume at least 10 μg/day and possibly up to 25 μg/day (Garabedian et al., 1991), and that most children aged 4 years and over, but maybe not those aged 1–3 years, had enough sun exposure for an adequate vitamin D synthesis. Thus, the SCF set a reference value of 10 μg/day for children 1–3 years, then ranges of 0–10 (4–10 years) and 0–15 (11–17 years) μg/day, the higher end of the ranges applying in case of minimal endogenous synthesis.

The UK has revised the DRVs for vitamin D (DH, 1991) (Section 4.1). There were insufficient data to set RNI for infants and children aged 0–3 years (SACN, 2016). 'Safe intakes'14 were set at 8.5–10 μg/day for all infants (including exclusively breastfed infants) and 10 μg/day for ages 1 to < 4 years. A RNI of 10 μg/day was set for subjects aged 4 years and over (Section 4.1).

An overview of DRVS for vitamin D for infants and children is presented in Table 2.

Table 2. Overview of dietary reference values for vitamin D for children

	SACN ( 2016 ) a ^a 'Safe intakes' for the age 0 to < 4 years, RNI afterwards.	D-A-CH ( 2015 ) b ^b AIs set considering a lack of endogenous vitamin D synthesis. Vitamin D supplementation of infants, and of children aged up to 2 years during winter, was recommended.	NCM ( 2014 ) c ^c PRI assuming some endogenous vitamin D synthesis.	NL (2012)d ^d AIs in case of lack of endogenous synthesis. Vitamin D supplementation (10 μg/day) of young children was recommended. ^,k ^k NL, Health Council of the Netherlands	IOM ( 2011 )	WHO/FAO ( 2004 ) e ^e PRI in case of no endogenous vitamin D synthesis.	Afssa ( 2001 )	SCF ( 1993 )	DH ( 1991 ) f ^f DRVs to be met by supplementation up to at least 2 years of age. DRVs revised by SACN (2016).
Age (months)	0–11	0 to < 12	6–11	0 to < 12	0 to < 12	0 to < 12	0 to < 12	6–11	7 to < 12
DRV (μg/day)	8.5–10	10	10	10	10g ^g AI.	5	20–25h ^h Based on the summary table of Afssa (2001). Supplementation of infants (10–20 μg/day), of children (18 months–5 years) during winter (10–20 μg/day), and of adolescents during winter and with low sun exposure was advisable.	10–25	7
Age (years)	1–17	1–18	1–18	1–18	1–18	1–18	1–3	1–3	1–3
DRV (μg/day)	10	20	10	10	15i ⁱ PRI considering minimal sun exposure.	5	10	10	7
Age (years)							4–19	4–10	4–18
DRV (μg/day)							5	0–10j ^j Acceptable ranges of intake. Zero in case of adequate endogenous synthesis, the higher end of the range in case of minimal endogenous synthesis.	0
Age (years)								11–17
DRV (μg/day)								0–15j ^j Acceptable ranges of intake. Zero in case of adequate endogenous synthesis, the higher end of the range in case of minimal endogenous synthesis.

^a 'Safe intakes' for the age 0 to < 4 years, RNI afterwards.
^b AIs set considering a lack of endogenous vitamin D synthesis. Vitamin D supplementation of infants, and of children aged up to 2 years during winter, was recommended.
^c PRI assuming some endogenous vitamin D synthesis.
^d AIs in case of lack of endogenous synthesis. Vitamin D supplementation (10 μg/day) of young children was recommended.
^e PRI in case of no endogenous vitamin D synthesis.
^f DRVs to be met by supplementation up to at least 2 years of age. DRVs revised by SACN (2016).
^g AI.
^h Based on the summary table of Afssa (2001). Supplementation of infants (10–20 μg/day), of children (18 months–5 years) during winter (10–20 μg/day), and of adolescents during winter and with low sun exposure was advisable.
ⁱ PRI considering minimal sun exposure.
^j Acceptable ranges of intake. Zero in case of adequate endogenous synthesis, the higher end of the range in case of minimal endogenous synthesis.
^k NL, Health Council of the Netherlands

4.3 Pregnancy and lactation

According to D-A-CH (2015), maternal serum 25(OH)D concentration influences that of the fetus (Hollis and Wagner, 2004b; Wagner et al., 2008b). The vitamin D concentration in breast milk can be influenced by intake (Hollis and Wagner, 2004a,b; Wagner et al., 2006) but with high doses up to 160 μg/day (Wagner et al., 2006; Hollis et al., 2011), which were not considered advisable by D-A-CH (Wagner et al., 2008a). The same AI as that for non-pregnant non-lactating women was thus set, i.e. 20 μg/day in case of lack of endogenous vitamin D synthesis.

The Nordic Council of Ministers (2014) considered the marked increase in serum 1,25(OH)₂D concentration during pregnancy, a correlation between maternal and neonatal vitamin D status (Markestad, 1983), and lower winter serum 25(OH)D concentrations in pregnant Nordic women (Bjorn Jensen et al., 2013; Brembeck et al., 2013). The Council also considered the 'normal' serum 25(OH)D concentrations in pregnant women supplemented with 10 μg/day vitamin D (Markestad et al., 1986), the improved vitamin D status at term by supplementation during pregnancy (Cranney et al., 2007; De-Regil et al., 2012; Lamberg-Allardt et al., 2013), and the limited data on health outcomes. Thus, the previous RI for pregnant or lactating women, i.e. 10 μg/day, was maintained.

The Health Council of the Netherlands (2012) advised vitamin D supplementation particularly for pregnant women with light skin and insufficient sun exposure, or those with dark skin (10 μg/day, maybe even prior to pregnancy) and noted the low vitamin D concentration in breast milk (IOM, 2011). The Council applied the same AI for pregnant or lactating women as for other young women.

IOM (2011) (Sections 5.1.2 and 5.1.3, Appendix B) found (i) insufficient evidence on the association between maternal serum 25(OH)D concentration and BMD during pregnancy, (ii) no effect of maternal 25(OH)D concentration in pregnancy on fetal calcium homeostasis or skeletal outcomes, (iii) negative skeletal outcomes in the newborn below the EAR-type value (40 nmol/L, Section 4.1) for maternal 25(OH)D concentration and (iv) no reduced skeletal BMC in children above the RDA-type value (50 nmol/L, Section 4.1) for maternal 25(OH)D concentration (Delvin et al., 1986; Javaid et al., 2006; Cranney et al., 2007; Viljakainen et al., 2010). The IOM also considered that neither maternal BMD nor maternal or fetal serum 25(OH)D concentrations could be used to set reference values for vitamin D during lactation. IOM (2011) noted that there is no evidence that the vitamin D requirement of lactating adolescents or women differs from that of non-lactating females in relation to maternal or child outcomes. Thus, the same EAR and RDA were set for pregnant or lactating women as for non-pregnant non-lactating women.

WHO/FAO (2004) considered the limited impact of changes in vitamin D metabolism during pregnancy on maternal requirements, the vitamin D transfer from mother to fetus, and the use of conventional prenatal vitamin D supplements to ensure adequate vitamin D status. The WHO/FAO estimated that there was no direct role for vitamin D in lactation because of the regulation of increased calcium needs by the PTH-related peptide (Sowers et al., 1996; Prentice, 1998) and the lack of evidence of any change in vitamin D metabolites during lactation (Kovacs and Kronenberg, 1997; Sowers et al., 1998). Vitamin D concentration in breast milk was considered as low (Specker et al., 1985), and the rare cases of nutritional rickets were almost always observed in breast-fed infants not exposed to the sun (Binet and Kooh, 1996; Brunvand and Nordshus, 1996; Gessner et al., 1997; Pettifor and Daniels, 1997). Evidence was lacking for an increased calcium or vitamin D transfer in milk after supplementation in lactating mothers (Sowers et al., 1998). Therefore, the same recommended nutrient intake of 5 μg/day was applied for pregnant and lactating women and for other younger women (19–50 years).

Afssa (2001) considered that pregnant women in France may have a deficient vitamin D status at the end of pregnancy, particularly in winter or early spring, even in the South of France. Vitamin D supplementation (25 μg/day during the last trimester, or a single dose of 5 mg at the seventh month) was also mentioned. The reference value of pregnant or lactating women was set at 10 μg/day.

The SCF (1993) considered that usual sun exposure in Europe may be insufficient to cover vitamin D needs, especially during the last trimester of pregnancy and at the end of winter, and that the ensuing vitamin D deficiency would affect mother and newborn (as neonatal vitamin D stores depend on maternal ones). The SCF (1993) recommended 10 μg/day to maintain 25(OH)D concentrations of pregnant and lactating women (Cockburn et al., 1980; Greer et al., 1981).

The UK has revised the DRVs for vitamin D (DH, 1991). The RNI of 10 μg/day proposed for subjects aged 4 years and over (Section 4.1) also applies to pregnant and lactating women (SACN, 2016).

An overview of DRVs for vitamin D for pregnant and lactating women is presented in Table 3.

Table 3. Overview of dietary reference values for vitamin D for pregnant and lactating women

	SACN (2016)a ^a PRI.	D-A-CH (2015)b ^b AI in case of lack of endogenous synthesis of vitamin D.	NCM (2014)c ^c PRI assuming some endogenous vitamin D synthesis.	IOM (2011)d ^d PRI considering minimal sun exposure.	NL (2012)b ^b AI in case of lack of endogenous synthesis of vitamin D. ^,g ^g NL, Health Council of the Netherlands	WHO/FAO (2004)e ^e PRI in case of no endogenous vitamin D synthesis.	Afssa (2001)	SCF (1993)a ^a PRI.	DH (1991)f ^f DRVs to be met by supplementation. DRVs revised by SACN (2016).
Pregnancy (μg/day)	10	20	10	15	10	5	10	10	10
Lactation (μg/day)	10	20	10	15	10	5	10	10	10

^a PRI.
^b AI in case of lack of endogenous synthesis of vitamin D.
^c PRI assuming some endogenous vitamin D synthesis.
^d PRI considering minimal sun exposure.
^e PRI in case of no endogenous vitamin D synthesis.
^f DRVs to be met by supplementation. DRVs revised by SACN (2016).
^g NL, Health Council of the Netherlands

5 Criteria (endpoints) on which to base dietary reference values

The Panel considered serum 25(OH)D concentration as a useful biomarker of vitamin D intake (in a population with low exposure to UV-B irradiation) and of vitamin D status in children and adults (Section 2.4.6). The Panel also considered that serum 25(OH)D concentration represents total vitamin D from exposure to both UV-irradiation (cutaneous synthesis) and dietary sources (Section 2.3.3). The Panel considered that the association between vitamin D intake and status for the purpose of deriving DRVs for vitamin D should be assessed under conditions of minimal endogenous vitamin D synthesis (Section 2.3.1). As indicated previously (Sections 2.4.1 and 4), there is an ongoing debate about the optimal range of serum 25(OH)D concentration and the cut-off values for defining deficiency, insufficiency and sufficiency.

Thus, the Panel reviewed data first on serum 25(OH)D concentration and health outcomes (Section 5.1), irrespective of the analytical method applied to measure serum 25(OH)D concentration (Section 2.4.1). Then, the Panel reviewed data on vitamin D intake (from supplements) and health outcomes (Section 5.2). Finally, the Panel reviewed and assessed data on the relationship between vitamin D intake (from food and supplements) and serum 25(OH)D concentration under conditions of minimal endogenous synthesis, and on factors potentially influencing this relationship (Section 5.3, Appendices C and D).

5.1 Serum 25(OH)D concentration and health outcomes

5.1.1 Serum concentration

The active metabolite 1,25(OH)₂D in association with VDR has a biological function not limited to bone, intestine, kidneys and parathyroid glands, but throughout the body, regulating many functions (Section 2.3.6). The Panel thus considered the relationships between vitamin D status, assessed by serum 25(OH)D concentration, and the risk of various health outcomes (musculoskeletal or non-musculoskeletal), to evaluate whether they might inform the setting of DRVs for vitamin D. This assessment was undertaken irrespectively of the analytical method applied to measure serum 25(OH)D concentration (Section 2.4.1).

The review of data on serum 25(OH)D concentration and musculoskeletal health outcomes in (healthy) adults and children is first described (Section 5.1.2). Then, the Panel reviewed data on serum 25(OH)D concentration and health outcomes in pregnancy (Section 5.1.3) and lactation (Section 5.1.4). Finally, an overview of available data on serum 25(OH)D and non-musculoskeletal health outcomes is given (Section 5.1.5).

For all of these outcomes, the Panel took as a starting point the results of the literature search and conclusions from the report by IOM (2011) (Section 4, Appendix B). This report by the IOM was based (i) on the systematic review (of RCTs (mainly), prospective cohort, case–control and before-after studies published in 1966–2006) by Cranney et al. (2007) on effectiveness and safety of vitamin D in relation to bone health, (ii) on another systematic review (of RCTs, non-randomised comparative studies, cohort and nested case–control studies and systematic reviews) by Chung et al. (2009) on vitamin D and/or calcium and various health outcomes, which focused, however, on RCTs published in 2006–2008 in relation to bone health outcomes to update the review by Cranney et al. (2007), and (iii) on additional literature search.
For all of these outcomes, the Panel also considered the report of the Agency for Healthcare Research and Quality (AHRQ) by Newberry et al. (2014), which is an update of Chung et al. (2009) for the period 2008–2013 with regard to data on vitamin D intake (and status) with or without calcium. The Panel considered as well the report by SACN (2016) submitted for public consultation in July–September 2015 and that served as a basis for updating the references values for vitamin D in the UK. The report by SACN (2016) took as a starting point the results of the literature search of the report by IOM (2011). It contains a review of human studies published up to 2014 in the document published for public consultation, further updated up to March 2016. For musculoskeletal health outcomes, the Panel also considered the systematic literature review (of systematic reviews (mainly) and RCTs published in 2000–2012) by Lamberg-Allardt et al. (2013) on vitamin D intake and status and health (including safety), which tried to identify a serum 25(OH)D concentration that would reflect sufficient vitamin D status and served as a basis for updating the reference values for vitamin D for the Nordic Nutrition Recommendations 2012 (Nordic Council of Ministers, 2014) (Section 4).
For its literature search related to musculoskeletal health outcomes in adults and children, as well as health outcomes in pregnancy and lactation, the Panel considered pertinent primary studies published from 2010 (after the IOM report) onwards until March 2015 in PubMed and/or as identified in Newberry et al. (2014) and/or SACN (2016), on the possible relationship between vitamin D status and health outcomes, with the aim to identify a serum 25(OH)D concentration to be used for deriving the DRVs for vitamin D (Also, using the same approach, the Panel considered pertinent primary studies on vitamin D intake and health outcomes, see Section 5.2).

Regarding the design of the primary studies considered, the Panel focused on intervention studies and prospective observational studies in healthy subjects, i.e. excluding cross-sectional studies (except for osteomalacia), case reports and ecological studies. The Panel notes that, in observational studies, positive, inverse, or lack of associations between 25(OH)D concentration and musculoskeletal health outcomes might be biased because of uncertainties in the methodology for measuring serum 25(OH)D concentration or confounded by factors that have not been properly addressed. In the following sections, for each musculoskeletal health outcome in adults and children, as well as each health outcome in pregnancy and lactation, first the intervention studies and then the prospective observational studies are described individually, and finally, an overall discussion and conclusion by health outcome is provided.

With the aim of setting DRVs for vitamin D, the Panel considered studies on vitamin D intake from food and/or daily or weekly supplementation using doses up to the UL for the respective population group (e.g. for adults: 100 μg/day) (EFSA NDA Panel, 2012a), and excluded studies reporting on lower frequency of consumption (e.g. monthly, once per trimester, or yearly administration).

5.1.2 Serum 25(OH)D concentration and musculoskeletal health outcomes

The Panel reviewed the available evidence on vitamin D status and musculoskeletal health outcomes, with the aim of identifying a target serum 25(OH)D concentration associated with low risk of compromised musculoskeletal health that can be used for the setting of DRVs for vitamin D.

The Panel considered musculoskeletal health outcomes to include BMD/BMC, risk of osteomalacia or of rickets, fracture risk, risk of falls/falling, muscle strength/muscle function/physical performance, and calcium absorption (Sections 2.2.1, 2.2.2.1 and 2.4.3). Markers of bone turnover (i.e. of bone formation and resorption) were not considered (Section 2.4.5).

The Panel decided to consider available data on bone measurements (BMC, BMD) in children and adults obtained via different techniques (e.g. dual-energy X-ray absorptiometry (DXA) or peripheral quantitative computed tomography (pQCT) (Appendix A) and after an appropriate study duration (e.g. at least 1 year (EFSA NDA Panel, 2012b)).

5.1.2.1 Adults

5.1.2.1.1 Bone mineral density/bone mineral content (BMD/BMC)

IOM ( 2011 ) (Section 4 and Appendix B) underlined that results from RCTs did not show an association between serum 25(OH)D concentration and BMD or bone loss. The IOM considered, however, that the majority of observational studies in post-menopausal women and older men supported an association between serum 25(OH)D concentration and BMD or change in BMD, particularly at the hip sites, and that 25(OH)D concentrations that were associated with an increase of bone loss at the hip ranged from < 30 to 80 nmol/L.

Lamberg-Allardt et al. (2013) based their conclusions about the possible relationship between 25(OH)D concentration and BMD or BMC in older adults on Cranney et al. (2007) and Chung et al. (2009) and their conclusions were in agreement with those derived by IOM (2011). Newberry et al. (2014) did not specifically report on the relationship between 25(OH)D concentration and BMC/BMD in adults beyond the conclusions of IOM (2011). With regard to bone health indices in adults aged 50 years and over, SACN (2016) additionally considered a systematic review by Reid et al. (2014) that included 23 studies (most of which were published between 1991 and 2009; four of the seven more recent studies were on patients or institutionalised subjects), two intervention studies (Kärkkäinen et al., 2010; Macdonald et al., 2013) and one prospective cohort study (Ensrud et al., 2009). However, no overall conclusion was drawn on the association between serum 25(OH)D concentration and risk for increase of bone loss.

The Panel retrieved 15 intervention and prospective observational studies in free-living adults, reporting on BMD/BMC in relation to 25(OH)D concentration and that were published after the report by IOM (2011). In the following section, the seven intervention studies and then the eight prospective observational studies are described individually. The results are then summarised, and an overall conclusion on BMD/BMC in adults is provided.

RCTs with vitamin D supplementation

In a double-blind 1-year RCT performed in Norway by Jorde et al. (2010), overweight men and women (21–70 years) received 500 μg vitamin D₃ per week (equivalent to 71 μg/day) (DP group n = 132), or placebo (PP group, n = 142). All subjects were given 500 mg/day calcium and 202 subjects completed the study. Mean (standard deviation, SD) serum 25(OH)D concentration increased from 58 (20) to 100 (20) nmol/L in the DP group and remained unchanged in the PP group (58 (20) nmol/L). After 1 year, there were no significant differences between the two groups regarding change in BMD (lumbar spine and hip). The Panel notes that raising mean 25(OH)D concentration from 58 to 100 nmol/L by weekly high dose supplementation with vitamin D for 1 year did not have an effect on BMD in these healthy overweight and mostly vitamin D sufficient subjects with an adequate calcium supply and who covered a large age range.

In a 1-year RCT by Islam et al. (2010), 200 apparently healthy young female factory workers (16–36 years) in Bangladesh received either: (1) daily 10 μg vitamin D15; (2) daily 10 μg vitamin D + 600 mg calcium; (3) daily 10 μg vitamin D and other micronutrients + 600 mg calcium; or (4) placebo. These women worked from dawn to dusk on all days of the week and wore concealing clothing (hands and faces uncovered). Mean 25(OH)D concentration was between 35 and 38 nmol/L among the groups at baseline, but was significantly (p < 0.001) higher in the three supplemented groups than in the placebo group (69 vs 36 nmol/L) at the end of the study. After adjustments for potential confounders, BMD and BMC increased significantly at the femoral neck (p < 0.001) and at the greater trochanter and Ward's triangle (for both, p < 0.05) in the supplemented groups compared with placebo, but there was no significant difference between groups at the lumbar spine (L2–L4). The Panel notes that raising mean 25(OH)D concentration from 35–38 nmol/L up to 69 nmol/L in these young Bangladeshi women with low sun exposure by vitamin D supplementation (with or without calcium) for 1 year was associated with a significant increase in BMD at the femoral neck, greater trochanter and Ward's triangle, but not at the lumbar spine.

In a randomly selected subsample of 593 subjects from a randomised population-based open trial with a 3-year follow-up in 3,432 women (aged 66–71 years) in Finland (Kärkkäinen et al., 2010), the intervention group (n = 287 completers) received daily 20 μg vitamin D₃ + 1,000 mg calcium for 3 years, while the control group (n = 306 completers) received neither supplementation nor placebo. The respective mean calcium intakes were 988 and 965 mg/day at baseline. The respective mean (SD) 25(OH)D concentrations were 50.1 (18.8) and 49.2 (17.7) nmol/L at baseline. At the end of the trial, serum 25(OH)D was significantly higher in the intervention group as compared to the control group (74.6 (21.9) vs 55.9 (21.8) nmol/L, p < 0.001). In the intention-to-treat (ITT) analysis, total body BMD increased significantly more in the intervention group than in the control group (0.84% vs 0.19%, p = 0.011) and the BMD decrease at Ward's triangle was lower in the intervention group (−2.69% vs −2.83%, p = 0.003). BMD changes at the lumbar spine, femoral neck, trochanter, and total proximal femur were not statistically different between groups. The women who were adherent (i.e. those who took at least 80% of their supplementation) showed significantly lower bone loss in femoral neck (−1.26% vs −1.73%, p = 0.002), Ward's triangle (−1.63% vs −2.83%, p < 0.0001), trochanter (0.25% vs –0.88%, p = 0.001), and total proximal femur (−0.84% vs −1.47%, p < 0.0001) than the women in the control group. Further, total body BMD increased more in the intervention group (1.31% vs 0.19%, p = 0.002). In contrast, the increase in lumbar spine BMD was lower in the intervention group than in the control group (0.67% vs 0.76%, p = 0.033). The Panel notes that raising mean 25(OH)D concentration from 50 nmol/L to 75 nmol/L by daily vitamin D and calcium supplementation for 3 years was associated with a significantly higher increase in total BMD in these women and, in subjects that adhered to the protocol, with a significantly lower bone loss in femoral neck, Ward's triangle, trochanter and total proximal femur, but a significantly lower increase in lumbar spine BMD compared to the control group. The Panel also notes that all analyses were unadjusted.

In an 18-month RCT with a factorial design in Australia by Kukuljan et al. (2011), 180 Caucasian men aged 50–79 years were randomised to: fortified milk (400 mL/day of milk containing 1,000 mg/day calcium and 20 μg/day vitamin D₃); exercise + fortified milk; exercise; or control (no milk, no exercise). Mean baseline serum 25(OH)D concentrations averaged 86.3 ± 36 nmol/L across the groups, in which no, one and 17 participants had serum 25(OH)D concentrations below 12.5 nmol/L, of 12.5–25 nmol/L and of 25–50 nmol/L, respectively. Serum 25(OH)D concentrations increased by an average of 21 nmol/L in the fortified milk compared with the two non-fortified milk groups after 12 months (p < 0.001), with no further increases observed at 18 months. Changes in BMD, bone structure, and strength at the lumbar spine, proximal femur (femoral neck), mid-femur, and mid-tibia were measured. There were no exercise-by-fortified milk interactions at any skeletal site. Main effect analysis showed that exercise led to a net gain in femoral neck section modulus (a measure for bending strength) and lumbar spine trabecular BMD, but there were no main effects of the fortified milk at any skeletal site. The Panel notes that raising mean 25(OH)D concentration from about 86 to 107 nmol/L by providing vitamin D₃ (with calcium) to these mostly replete men for 18 months did not enhance BMD. This suggests that other factors may confound the relationship between vitamin D intake, serum 25(OH)D and BMD or that, above a certain 25(OH)D concentration, there is no effect of additional calcium and vitamin D on BMD.

In a 2-year double-blind RCT in the US, Nieves et al. (2012) investigated the effect of 25 μg/day vitamin D₃ supplementation vs placebo on bone loss in post-menopausal African-American women (mean age about 62 years) (ITT: n = 103) and the influence of polymorphisms in the gene encoding VDR (Sections 2.2.1, 2.3.6 and 2.5). All women received calcium supplementation (to reach a total intake of 1,000 mg/day). Mean (± SD) baseline 25(OH)D concentrations were 29 ± 13 and 29 ± 14 nmol/L in the intervention (n = 55) and placebo (n = 48) groups, respectively, and in 50% of the subjects, 25(OH)D concentration was below 25 nmol/L. After 2 years, serum 25(OH)D significantly increased by 27.5 nmol/L in the intervention group (p < 0.001), but did not change in the placebo group. Two-year changes in spine or hip BMD did not significantly differ between groups at any skeletal site. When the entire population was divided according to Fok1 polymorphism (that has been associated with BMD in post-menopausal women), there were no significant differences in the 25(OH)D response to vitamin D supplementation by genotype. Despite similar elevations in 25(OH)D, femoral neck BMD was only responsive to vitamin D supplementation in FF subjects (n = 47), not in Ff/ff subjects (n = 31). The Panel notes that, in these post-menopausal African American women, raising mean 25(OH)D concentration from about 29 to 56 nmol/L by vitamin D supplementation was not associated with significantly different 2-year changes in spine or hip BMD compared with the placebo group, both groups having a mean baseline 25(OH)D concentration of 29 nmol/L and sufficient calcium supply. The Panel also notes that the possible relationship between baseline or follow-up 25(OH)D concentration and BMD may depend, among other factors, on genetic predisposition. In this context, the Panel notes that, with regard to the Fok1 polymorphism, the reported frequency of the FF genotype among various populations was reported to be between 40% and 50% (Laaksonen et al., 2004; Sanwalka et al., 2013).

In a 1-year double-blind RCT in Scotland, Macdonald et al. (2013) determined whether daily vitamin D₃ supplementation compared with placebo affects BMD change in healthy Caucasian post-menopausal women aged 60–70 years (ITT: n = 264). Mean intakes of calcium and vitamin D from food and other supplements amounted, respectively, to around 1.3 g/day and 5 μg/day at baseline in all groups. Total mean vitamin D intake (i.e. with food and all supplements) amounted to about 5, 15, and 30 μg/day in the placebo (n = 90), 10 μg supplemented (n = 84) and 25 μg supplemented (n = 90) groups, respectively. Mean (± SD) baseline 25(OH)D was 33.8 ± 14.6 nmol/L. The 25(OH)D changes were −4.1 ± 11.5 nmol/L, +31.6 ± 19.8 nmol/L, and +42.6 ± 18.9 nmol/L in the placebo, 10 μg, and 25 μg groups, respectively. After adjustments for potential confounders, mean BMD loss at the hip, but not lumbar spine, was significantly less for the 25 μg vitamin D group (0.05 ± 1.46%) compared with the 10 μg vitamin D or placebo groups (0.57 ± 1.33% and 0.60 ± 1.67%, respectively) (p< 0.05). Neither at baseline nor at the final visit were significant associations between serum 25(OH)D and mean total hip BMD or lumbar spine BMD found. The Panel notes that raising mean 25(OH)D concentration from about 34 to 65 and 76 nmol/L, respectively, by supplemental doses of 10 or 25 μg/day of vitamin D for 1 year in these post-menopausal women did not result in corresponding dose-dependent effects on BMD. This suggests that other factors may confound the relationship between vitamin D intake, serum 25(OH)D and BMD, and that 25(OH)D concentrations above 34 nmol/L were not associated with BMD in this study.

In a 2-year double-blind RCT in 409 home-dwelling women (70–80 years) in Finland, Uusi-Rasi et al. (2015) (Sections 5.1.2.1.4, 5.1.2.1.5, 5.2.3 and 5.2.4) investigated the effect of daily vitamin D₃ supplementation (20 μg) with or without exercise on BMD assessed as secondary outcome. The subjects were randomly assigned to four groups, two vitamin D groups who received daily vitamin D₃ supplementation, either with (n = 103) or without exercise (n = 102), and two placebo groups, either with (n = 103) and without exercise (n = 102). Mean calcium and vitamin D intakes at baseline and at the end of the study were 1,098 and 1,212 mg/day, respectively, for calcium, and 10.4 and 10.5 μg/day, respectively, for vitamin D. The mean (SD) baseline serum 25(OH)D concentrations among groups ranged between 65.5 (17.5) and 69.5 (18.0) nmol/L (not significantly different). After 2 years, while serum 25(OH)D concentrations remained stable in the placebo groups, concentrations increased in the vitamin D groups to 92.5 (18.5) nmol/L. ITT analyses yielded that femoral neck BMD declined in all groups, with the mean decline being greatest in the placebo without exercise group and differed significantly from that in all other study groups (p < 0.02, p < 0.01, p < 0.04). Lumbar spine BMD did not change significantly in any group, whereas trabecular bone density at the tibia was slightly but significantly increased in the vitamin D and exercise group compared to the placebo without exercise group (p = 0.02). However, there was no difference between vitamin D without exercise group and the placebo without exercise group. Also, in none of the intervention groups, changes in either tibial shaft cortical or area density differed from the placebo without exercise group. The Panel notes that, in this study, vitamin D supplementation of 20 μg/day (above the usual intake of 10 μg/day) leading to an increase of mean serum 25(OH)D concentration from about 66 nmol/L up to 93 nmol/L, resulted in a lower decrease in femoral neck BMD as compared to placebo. The Panel also notes that similar effects were observed with exercise without vitamin D or with exercise and vitamin D and that vitamin D alone had no effect on BMD at other sites.

Prospective observational studies

Bolland et al. (2010) examined the association between baseline serum 25(OH)D concentration and multiple health outcomes (Sections 5.1.2.1.3, 5.1.2.1.4 and 5.1.2.1.5) in 1,471 community-dwelling women (mean age 74 years) who took part in a 5-year calcium supplementation study in Australia. Fifty per cent of women had a seasonally adjusted 25(OH)D concentration < 50 nmol/L and these women were significantly older, heavier, and less physically active and had more comorbidities than women with a seasonally adjusted 25(OH)D concentration ≥ 50 nmol/L. After adjustments for potential confounders (including treatment allocation to calcium or placebo), women with a seasonally adjusted baseline 25(OH)D concentration < 50 nmol/L and those with 25(OH)D concentrations ≥ 50 nmol/L did not show any difference in change in bone density (lumbar spine, total femur, total body). The Panel notes that this study of community-dwelling older women showed no difference in BMD change in those with a seasonally adjusted 25(OH)D concentration < 50 nmol/L compared with those with 25(OH)D concentrations ≥ 50 nmol/L over a 5-year period.

In a cohort of 1,097 healthy peri- or post-menopausal Caucasian Danish women (45–57 years, median: 51 years) with a 16-year follow-up, Rejnmark et al. (2011) (Section 5.1.2.1.3) investigated the association of tertiles of PTH concentrations (upper tertile ≥ 4.5 pmol/L) with BMD (assessed at the 10-year follow-up) stratified according to baseline 25(OH)D concentrations < 50 nmol/L, at 50–80 nmol/L, or > 80 nmol/L, after adjustments for potential confounders. Mean baseline plasma 25(OH)D was 65 ± 31 nmol/L. Within the group of women with plasma 25(OH)D < 50 nmol/L at baseline, high (≥ 4.5 pmol/L) PTH concentration, compared to low PTH concentration (< 4.5 pmol/L), was associated with a significantly larger decrease in lumbar spine BMD between baseline and the 10-year visit (−5.6 ± 7.0% vs −3.4 ± 7.0%, p = 0.01) after adjustments for potential confounders. In contrast, high (vs low) PTH concentration was not associated with bone loss rates at the lumbar spine in women with 25(OH)D concentrations of 50–80 nmol/L or in women with 25(OH)D concentrations > 80 nmol/L. However, there was no influence of plasma 25(OH)D concentration on the relationships of PTH with 10-year changes in BMD at the total hip, femoral neck, and whole body. The Panel notes that this study indicates that, in these women, a greater 10-year BMD loss at the lumbar spine was associated with a baseline plasma 25(OH)D concentration < 50 nmol/L at higher PTH concentration and that the relationship between 25(OH)D concentration and BMD depends on PTH.

In a cohort of mobile community-dwelling Chinese men aged 65 years and over (n = 712) with a 4-year follow-up, Chan et al. (2011) (Section 5.1.2.1.3) examined serum 25(OH)D in relation to BMD. Mean baseline 25(OH)D concentration was 78.2 ± 20.5 nmol/L, and, respectively, 5.9%, 41.5%, and 52.6% had concentration below 50 nmol/L, of 50 to < 75 nmol/L, or 75 nmol/L or higher. After adjustments for potential confounders, there was no association between serum 25(OH)D concentration and 4-year percentage change in BMD at total hip, spine, and femoral neck. The results remained unchanged when subjects were divided into quartiles of serum 25(OH)D, i.e. concentration of the first quartile ≤ 63 nmol/L vs concentration > 63 nmol/L. The Panel notes that, in this study in men with a mean serum 25(OH)D concentration of about 78 nmol/L at baseline, no association was found between baseline serum 25(OH)D concentration (continuous variable or over quartiles of < 63 nmol/L up to > 91 nmol/L) and a lower 4-year bone loss at any site.

In a cohort study among 2,614 community-dwelling white and black women and men aged ≥ 70 years in the USA, secondary analyses were conducted by Barbour et al. (2012) (Section 5.1.2.1.3) to determine the average annual change in hip areal BMD (aBMD) by quartiles of 25(OH)D concentration (< 44.5 nmol/L, 44.5–61 nmol/L, 61–79.8 nmol/L, > 79.8 nmol/L; mean baseline value was not reported). Blood samples were drawn at year 2, which formed the baseline for this analysis, and hip aBMD was measured at baseline, years 3, 5 or 6, 8, and 10. After adjustments for potential confounders, lower 25(OH)D was associated with greater aBMD loss (p trend = 0.024). Participants in the top 25(OH)D quartile had significantly lower annualised hip aBMD loss (−0.55%, 95% confidence interval (CI): −0.48 to −0.62%) compared with those in the lowest quartile (−0.65%, 95% CI: −0.58 to −0.72%). The Panel notes that, in this study, a baseline serum 25(OH)D concentration below 44.5 nmol/L (lowest quartile) was associated with a 0.1% higher annual hip aBMD loss compared to serum 25(OH)D > 79.8 nmol/L.

In a case-cohort study with a 4.6 year follow-up in the US, Barrett-Connor et al. (2012) (Section 5.1.2.1.3) tested the hypothesis that combinations of 'low' serum 25(OH)D concentration (< 50 nmol/L), 'low' sex hormones (SH) (bioavailable testosterone (BioT) < 163 ng/dL; bioavailable estradiol (BioE) < 11 pg/mL), and 'high' sex hormone binding globulin (SHBG) (> 59 nmol/L) would have a synergistic effect on total hip BMD loss. Participants were a random subsample of 1,468 men (mean age: 74 years) from a larger prospective cohort study plus 278 men from this cohort with incident non-spine fractures. One quarter of the men had baseline 25(OH)D < 50 nmol/L (mean 38.8 nmol/L). After adjustments for potential confounders, 'low' 25(OH)D in isolation, and 'low' BioT with or without 'low' 25(OH)D, were not significantly related to BMD loss. However, the combination of 25(OH)D < 50 nmol/L with 'low' BioE and/or 'high' SHBG was associated with significantly lower baseline total hip BMD (p = 0.03, p = 0.002) and higher annualised rates of hip bone loss (p = 0.007, p = 0.0006), than SH abnormalities alone or no abnormality. The Panel notes that the adverse effect of 'low' BioE and/or 'high' SHBG serum concentrations on total hip BMD was more pronounced in older men with baseline serum 25(OH)D concentration < 50 nmol/L (lowest quartile, mean 38.8 nmol/L), whereas 25(OH)D concentration < 50 nmol/L in isolation was not associated with BMD.

In a population-based cohort of 192 apparently healthy ambulatory older Lebanese men (n = 64) and women (n = 128) aged 65–85 years, with a median 4-year follow-up, Arabi et al. (2012) analysed the association of 25(OH)D, PTH and body composition with change in BMD at the lumbar spine, hip (femoral neck, trochanter, total hip), and forearm and subtotal body BMC. For serum 25(OH)D and PTH, average of baseline and follow-up concentrations were used in the analyses. Mean 25(OH)D concentration was 36.8 ± 16 nmol/L and BMD significantly decreased at all skeletal sites except at the spine. Multivariate analyses of per cent changes in BMD (at all skeletal sites) or subtotal body BMC showed that 25(OH)D was not a significant predictor, contrary to changes in body composition and PTH. The Panel notes that this study showed no association between serum 25(OH)D and 4-year bone loss at the lumbar spine, hip or forearm in a population with a mean serum 25(OH)D concentration of about 37 nmol/L (average of baseline and follow-up).

In a cohort study in Japan, Kitamura et al. (2013) explored the association between serum 25(OH)D concentration, PTH concentration and 5-year changes in BMD of the lumbar spine and femoral neck in 482 independently living post-menopausal women (mean age, range: 63.1 years, 55–74 years). Their mean baseline serum 25(OH)D concentration was 56 nmol/L. In the serum 25(OH)D quartiles (< 46.5, 46.5 to < 56.1, 56.1 to < 65.1, ≥ 65.1 nmol/L), mean concentrations were 37.5 ± 7.5, 51.2 ± 2.8, 60.3 ± 2.4, and 74.7 ± 7.7 nmol/L, respectively. Mean calcium intake was not significantly different between serum 25(OH)D quartiles (519–536 mg/day). After adjustment for potential confounders, there was no significant association between baseline serum 25(OH)D concentration (as quartiles) and change in BMD (at either site). The Panel notes that this study indicates that, even at a rather low calcium intake, the lowest baseline quartile serum 25(OH)D concentration (< 46.5 nmol/L, mean of about 38 nmol/L) was not associated with a higher 5-year post-menopausal bone loss at the lumbar spine or femoral neck.

In a cohort of 922 women during the menopausal transition (mean age 48.5 ± 2.7 years) at five US clinical centres and with an average follow-up of 9.5 years, Cauley et al. (2015) (Section 5.1.2.1.3) determined if higher serum 25(OH)D baseline concentration is associated with slower loss of BMD. BMD was measured at each annual visit. The mean 25(OH)D concentration was 54.5 nmol/L; 43% of the women had 25(OH)D concentrations < 50 nmol/L. Changes in lumbar spine and femoral neck BMD across menopause were not significantly associated with serum 25(OH)D concentration. The Panel notes that, in this study, baseline serum 25(OH)D concentration (mean 54.5 nmol/L) was not associated with changes in lumbar spine and femoral neck BMD across menopause.

Conclusions on BMD/BMC in adults

Among the 15 studies identified, most of which were in older free-living adults, the Panel notes the heterogeneity of study designs, populations and skeletal sites investigated. The Panel considers that the sensitivity of serum concentrations of 25(OH)D in predicting losses in BMD/BMC may be limited because of confounding by a variety of factors (e.g. PTH, genetic factors, sex steroids, body composition, age, sex, calcium intake, life-style factors, baseline values, season of assessment, and possible other yet unknown factors) that have only been partly considered in these analyses. Furthermore, observational studies mostly used single measurements of 25(OH)D concentration, thus possible long-term changes in 25(OH)D concentration were not considered in the analyses of the relationship with BMD/BMC changes.

Of the seven RCTs with vitamin D supplementation durations between 1 and 3 years, two RCTs in women indicated that daily vitamin D and calcium supplementation that led to an increase in mean 25(OH)D concentrations from 35–38 nmol/L to 69 nmol/L (Islam et al., 2010) and from 50 nmol/L to 75 nmol/L (Kärkkäinen et al., 2010), respectively, was associated with a significantly higher increase in BMD compared to the control group. In subjects that adhered to the protocol, raising mean 25(OH)D concentration from 50 nmol/L to 75 nmol/L was also associated with a significantly lower bone loss in femoral neck, Ward's triangle, trochanter and total proximal femur (Kärkkäinen et al., 2010). In one RCT, vitamin D supplementation alone, which led to an increase of mean serum 25(OH)D concentration from about 66 nmol/L up to 93 nmol/L, resulted in a lower decrease in femoral neck BMD as compared to placebo, but had no effect on BMD at other sites (Uusi-Rasi et al., 2015). However, in four RCTs, an increase in serum 25(OH)D concentration from a mean of 29 nmol/L (Nieves et al., 2012), 34 nmol/L (Macdonald et al., 2013), 58 nmol/L (Jorde et al., 2010) and 86 nmol/L (Kukuljan et al., 2011) up to 56 nmol/L, 76 nmol/L, 100 nmol/L and 107 nmol/L, respectively, after vitamin D supplementation or consumption of vitamin D-fortified food (with or without calcium), did not result in a change in BMD.

Of the eight prospective observational studies, one reported a 0.1% higher annual hip aBMD loss associated with baseline 25(OH)D concentrations < 45 nmol/L (lowest quartile), as compared to 25(OH)D concentrations above 80 nmol/L (highest quartile) (Barbour et al., 2012). One study found a significant relationship between PTH concentration and 10-year BMD loss at the lumbar spine at baseline serum 25(OH)D concentrations of < 50 nmol/L (Rejnmark et al., 2011). A third study observed an association between annual hip BMD loss and baseline 25(OH)D concentrations < 50 nmol/L (lowest quartile, mean 39 nmol/L) only in subjects with 'low' sex steroid concentrations (Barrett-Connor et al., 2012). However, three studies found no difference in (4- or 5-year) BMD changes at any sites between baseline serum 25(OH)D concentrations in the lowest quartile (< 46.5 nmol/L, (Kitamura et al., 2013); < 50 nmol/L (Bolland et al., 2010); < 63 nmol/L, (Chan et al., 2011)) and higher concentrations. Two other studies also did not find an association between BMD or BMC losses and serum concentrations of 25(OH)D in populations with mean 25(OH)D of 37 nmol/L (average of baseline and 4-year-follow-up) (Arabi et al., 2012) or 55 nmol/L (baseline) (Cauley et al., 2015).

The Panel notes that two RCTs (Islam et al., 2010; Kärkkäinen et al., 2010) indicate that BMD may increase when mean serum 25(OH)D concentration increases from about 35–38 to 69 nmol/L in young women and from 50 to 75 nmol/L in older women and that BMD losses at subsites may be less pronounced when mean serum 25(OH)D concentration is increased from about 50 to 75 nmol/L in these older women. The Panel also notes that three observational studies (Rejnmark et al., 2011; Barbour et al., 2012; Barrett-Connor et al., 2012) suggest that baseline serum 25(OH)D concentration below 45–50 nmol/L (alone (Barbour et al., 2012) or in combination with high PTH concentration or low' BioE and/or 'high' SHBG (Rejnmark et al., 2011; Barrett-Connor et al., 2012)) may be associated with increased BMD losses at various sites. However, the Panel considers that the majority of both RCTs and observational studies do not report increased BMD/BMC losses at or below similar serum 25(OH)D concentrations (baseline mean or lowest quartile). The Panel notes that other factors can interfere with the association between 25(OH)D and BMD/BMC and thus may contribute to these inconsistencies. The Panel concludes that, altogether, these 15 studies in apparently healthy adults, published after the report by IOM (2011), do not provide sufficient evidence for a conclusion on a serum 25(OH)D concentration below which there is an increased risk of BMD/BMC loss.

The IOM had considered that results from RCTs did not show an association between serum 25(OH)D concentration and BMD or bone loss, but that the majority of observational studies in post-menopausal women and older men supported an association between serum 25(OH)D concentration and BMD or change in BMD, particularly at the hip sites. IOM also considered that serum 25(OH)D concentrations that were associated with an increase in bone loss at the hip ranged from below 30 to 80 nmol/L. Taking into account the conclusions of IOM ( 2011 ) and the studies published thereafter, the Panel considers that there is some evidence, mostly from observational studies in mainly older adults, that the risk of increased BMD/BMC loss in free-living adults is higher with a serum 25(OH)D concentration below 50 nmol/L.

5.1.2.1.2 Osteomalacia

Only one study (Priemel et al., 2010), considered by IOM ( 2011 ), in 675 subjects aged 20–100 years (mean age = 58.7 years in males (n = 401) and 68.3 years in females (n = 274)), provides information on serum 25(OH)D concentration and osteomalacia (Section 2.2.2.1) assessed by post mortem bone biopsies. These subjects had been residing in Germany and died for reasons not related to cancer, metabolic disorders, or bone diseases. Priemel et al. (2010) assessed bone undermineralisation by pathological accumulation of osteoid, and defined osteomalacia as a ratio of osteoid volume (OV, i.e. bone matrix that is not mineralised) to total bone volume (BV) greater or equal to 2%. Only a few subjects had osteomalacia (OV/BV ≥ 2%) at serum 25(OH)D concentration above 50 nmol/L and no subject had osteomalacia at serum concentration of at least 75 nmol/L. By further inspecting the graphical presentation of the results of this study, IOM (2011) (Section 4. and Appendix B) noted that about 1% of subjects with a serum 25(OH)D concentration above 50 nmol/L had osteomalacia, while less than half of the subjects with serum 25(OH)D concentrations below 40 or even 25 nmol/L had osteomalacia. IOM (2011) used this study to consider that a serum 25(OH)D concentration of 50 nmol/L provides coverage for at least 97.5% of the population. The Panel notes that some concerns with regard to limitations of the Priemel study have been raised, such as the histomorphometric threshold used to define osteomalacia and the validity of post mortem 25(OH)D measurements (Aspray and Francis, 2013). However, the Panel considers that the threshold of OV/BV ≥ 2% used to define osteomalacia by Priemel et al. (2010) is a conservative approach. The Panel also notes that no studies are available showing whether post mortem 25(OH)D measurements are valid.

Lamberg-Allardt et al. (2013) referred to the conclusion of IOM (2011) regarding osteomalacia and stated that no additional reduction in the risk of osteomalacia is to be expected at serum 25(OH)D concentration above 50 nmol/L. Newberry et al. (2014) also did not address the relationship between 25(OH)D and osteomalacia beyond the report by IOM (2011). SACN (2016) considered two cross-sectional studies (Preece et al., 1975; Gifre et al., 2011), as well as case reports, on patients with osteomalacia from early 1940s to 2013 and concluded that evidence on vitamin D and osteomalacia is limited and drawn mainly from case reports, that there is no clear serum 25(OH)D threshold concentration below which the risk of osteomalacia is increased, but noted that mean concentrations (in patients) were below about 20 nmol/L in all the studies considered. The Panel did not retrieve any additional pertinent primary study published from 2010 onwards.

The Panel notes that no recently published relevant data from RCTs or prospective observational studies on the association between serum 25(OH)D concentration and ostemalacia are available.

The Panel takes into account the findings by SACN (2016), based mainly on case-reports and two cross-sectional studies in patients with overt osteomalacia at mean serum 25(OH)D concentrations below about 20 nmol/L. Based on the evidence available (Priemel et al., 2010) and in line with the conclusion of IOM (2011), the Panel considers that the risk of vitamin D-deficiency osteomalacia appears to be small with serum 25(OH)D concentration at or above 50 nmol/L.

5.1.2.1.3 Fracture risk

IOM ( 2011 ) (Section 4 and Appendix B) reported that there was a wide variation in serum 25(OH)D concentrations below which fracture risk may be increased and that this was observed for concentrations between 30 and 70 nmol/L.

Lamberg-Allardt et al. (2013) based their conclusions about risk of fractures in older adults on three systematic reviews (Avenell et al., 2009; Chung et al., 2009; Vestergaard et al., 2011). The overall conclusion in the NNR 2012 is that intervention with vitamin D alone has not been proven effective in preventing fractures in older adults. Moreover, Lamberg-Allardt et al. (2013) concluded that, although a threshold for serum 25(OH)D concentration of 74 nmol/L was considered to show a reduction in total fracture incidence, the variability in analytical methods and the fact that serum 25(OH)D was assayed only in subsamples, made this threshold unreliable. Newberry et al. (2014) did not identify any recent RCTs that assessed the effect of interventions of vitamin D alone on fracture risk. They reported on six recent observational studies that assessed the association between serum 25(OH)D and fracture risk (Cauley et al., 2011; Barbour et al., 2012; Barrett-Connor et al., 2012; de Boer et al., 2012; Holvik et al., 2013; Looker, 2013) and concluded that results were inconsistent among them. SACN (2016) additionally reported that evidence from five studies (Cauley et al., 2010, 2011; Nakamura et al., 2011; Barbour et al., 2012; Rouzi et al., 2012) is mixed. SACN (2016) also considered studies (intervention and cohorts studies, systematic review of observational studies) about prevention of stress fractures in younger adults (less than 50 years) that were military personnel. Such a population was not considered by the Panel in this section (with the aim of setting DRVs for vitamin D for the general population).

The Panel retrieved 15 relevant prospective observational studies in free-living adults (but no RCTs), reporting on fractures in relation to 25(OH)D concentration and that were published after the report by IOM (2011). In the following section, the 15 prospective observational studies are described individually. The results are then summarised, and an overall conclusion on fracture risk is provided.

Prospective observational studies

In a case-cohort study in men aged 65 years and older, Cauley et al. (2010) followed 436 men with incident non-spine fractures, including 81 hip fractures, and a random subcohort of 1,608 men over an average of 5.3 years. The mean baseline total 25(OH)D concentration was 61.5 ± 19.5 nmol/L in non-spine fracture subjects, 53.8 ± 19.8 nmol/L in hip fracture subjects and 63.0 ± 19.5 nmol/L in controls (non-spine fracture subjects versus non-patients, p = 0.14; hip fracture subjects versus controls, p < 0.0001). Serum 25(OH)D concentration was unrelated to non-spine fractures. Compared with men in the top quartile of total 25(OH)D concentration (≥ 70 nmol/L), the hazard ratio (HR) of hip fracture was 2.36 (95% CI: 1.08–5.15) for men in the lowest quartile (< 50 nmol/L) (p = 0.009 for trend), after adjustments for potential confounders.16 The results were not always statistically significant when other additional adjustments were considered.17 The Panel notes that, in these older men, serum 25(OH)D concentration < 50 nmol/L (lowest quartile) was associated with an increased risk for hip, but not for non-spine fractures.

Bolland et al. (2010) examined the association between baseline serum 25(OH)D concentration and multiple health outcomes (Sections 5.1.2.1.1, 5.1.2.1.4 and 5.1.2.1.5) in 1,471 community-dwelling women (mean age 74 years) who took part in a 5-year calcium supplementation study in Australia. Fifty per cent of women had a seasonally adjusted 25(OH)D concentration < 50 nmol/L. After adjustments for potential confounders (including treatment allocation to calcium or placebo), women with a seasonally adjusted baseline 25(OH)D concentration < 50 nmol/L were not at increased risk of fracture (hip, vertebral, distal forearm, or osteoporotic), compared with those with 25(OH)D concentration ≥ 50 nmol/L. The Panel notes that this study of community-dwelling older women with a seasonally adjusted 25(OH)D concentration < 50 nmol/L compared with those with 25(OH)D concentration ≥ 50 nmol/L showed no increased risk of fractures over a 5-year period.

In a nested case–control study in the USA in 400 white, 381 black, 193 Hispanic, 113 Asian and 46 Native American women (aged 50–79 years), Cauley et al. (2011) evaluated the incidence of fractures (all types) over an average of 8.6 years. In multivariable models, compared with concentrations < 50 nmol/L, higher baseline 25(OH)D concentrations ≥ 75 nmol/L were associated with a lower risk of fracture in white women (for 50 to < 75 nmol/L, odds ratio (OR): 0.82; 95% CI: 0.58–1.16; for ≥ 75 nmol/L: OR: 0.56; 95% CI: 0.35–0.90, p trend = 0.02). In contrast, higher 25(OH)D (≥ 50 nmol/L) compared with concentrations < 50 nmol/L were associated with a higher risk of fracture in black women (OR: 1.45; 95% CI: 1.06–1.98, p trend = 0.043), after adjustment for potential confounders. In Asian women, the OR for fracture at higher 25(OH)D concentrations (≥ 75 nmol/L) compared with 25(OH)D < 50 nmol/L, was 2.78 (95% CI: 0.99–7.80, p trend = 0.04). There was no association between 25(OH)D and fracture in Hispanic or Native American women. The Panel notes that, in this study, associations between 25(OH)D and fracture by race/ethnicity were divergent and that serum 25(OH)D were associated with significantly lower fracture risk in white women with baseline concentrations ≥ 75 nmol/L, but a higher fracture risk in black women with baseline concentrations ≥ 50 nmol/L.

In a cohort study, Nakamura et al. (2011) followed-up 773 community-dwelling Japanese women aged 69 years and older, for 6 years. Mean serum 25(OH)D concentration was 60.0 ± 17.6 nmol/L and mean calcium intake was 586 ± 259 mg/day. The adjusted HRs of limb and vertebral fracture for the first quartile (< 47.7 nmol/L) and the third quartile (59.2–70.9 nmol/L) of baseline serum 25(OH)D concentration, compared to the fourth quartile (≥ 71.0 nmol/L), were 2.82 (95% CI: 1.09–7.34) and 2.82 (95% CI: 1.09–7.27), respectively.18 The pooled adjusted HR was 0.42 (95% CI: 0.18–0.99) when the incidence in the fourth quartile (≥ 71.0 nmol/L) was compared to the other three quartiles combined (< 71.0 nmol/L). The Panel notes that, in this study in Japanese women with rather low calcium intake, risk for limb and vertebral fracture was higher at baseline serum 25(OH)D concentration < 71 nmol/L (quartiles Q1–Q3).

In a cohort study, Robinson-Cohen et al. (2011) followed-up 2,294 US Caucasian and African American men and women (mean age: 74 years) for a median duration of 13 years. Baseline serum 25(OH)D was below 37.5 nmol/L for 382 participants. After adjustments for potential confounders, serum 25(OH)D concentration less than 37.5 nmol/L was associated with a 61% greater risk of hip fracture (95% CI: 12–132%). The Panel notes that this study in both Caucasian and African American subjects indicated a greater risk for hip fractures at baseline serum 25(OH)D concentration < 38 nmol/L.

In a cohort study in Danish women (median age: 51 years) followed-up for 16 years (assessment after 10 years of follow-up) and with a mean baseline plasma 25(OH)D of about 65 nmol/L (Section 5.1.2.1.1), Rejnmark et al. (2011) also examined the risk of (all) fractures according to plasma 25(OH)D (below 50 nmol/L, at 50–80 nmol/L, and above 80 nmol/L) and tertiles of PTH concentrations. Plasma 25(OH)D concentrations per se were not associated with the risk of any fracture. High PTH concentrations (> 4.5 pmol/L) were associated with an increased fracture risk at 25(OH)D concentrations < 50 nmol/L (HR_adj = 1.71; 95% CI: 1.1–2.66, p < 0.01) and at 25(OH)D concentrations 50–80 nmol/L (HR_adj = 1.60; 95% CI: 1.07–2.37, p < 0.02). The Panel notes that this study in women indicated that baseline plasma 25(OH)D concentrations per se were not associated with fracture risk, but were related to fracture risk at concentrations < 80 nmol/L at high PTH concentrations. Thus, the relationship between 25(OH)D concentration and fracture risk was shown to depend on PTH.

In a cohort study in mobile community-dwelling Chinese men aged at least 65 years whose mean baseline 25(OH)D concentration was about 78 ± 20 nmol/L (Section 5.1.2.1.1), Chan et al. (2011) also found, in multivariate regression analyses, no association between baseline serum 25(OH)D concentration (continuous variable or over quartiles of < 63 nmol/L up to > 91 nmol/L) and the 4-year risk of non-vertebral or hip fractures. The Panel notes that this study in men with a mean serum 25(OH)D concentration of about 78 nmol/L found no association between baseline serum 25(OH)D concentrations and risk of non-vertebral or hip fractures.

In a cohort study with a median follow-up time of 6.4 years in US community-dwelling white and black men and women aged ≥ 70 years (Section 5.1.2.1.1), Barbour et al. (2012) also investigated whether increasing serum 25(OH)D and decreasing PTH concentrations are associated with decreased risk of hip and any non-spine fracture, assessed every 6 months after year 2 ('baseline'). In multivariate analyses, there was no significant association between the risk of hip fracture and 25(OH)D concentration assessed as quartiles (≤ 44.5 nmol/L, 44.5–60.9 nmol/L, 60.9–79.9 nmol/L, compared to > 79.9 nmol/L). The Panel notes that this study in older subjects found no evidence of an association between baseline serum 25(OH)D concentration ranging from < 45 nmol/L to ≥ 80 nmol/L (extreme quartiles) and any non-spine fractures.

In a case-cohort study in older men (mean age: 74 years) in the USA, of which one quarter had 25(OH)D concentration < 50 nmol/L with a mean of 38.8 nmol/L, Barrett-Connor et al. (2012) (Section 5.1.2.1.1) also tested the hypothesis that combinations of low 25(OH)D (< 50 nmol/L), low SH, and high SHBG would have a synergistic effect on non-spine fracture risk. Compared to men with 25(OH)D > 50 nmol/L, BioT > 163 ng/dL, BioE > 11 pg/mL, SHBG < 59 nmol/L, multivariate analyses showed no significant association between risk for incident non-spine fractures and low 25(OH)D (< 50 nmol/L) in isolation, or low BioE and/or high SHBG in isolation. The multivariate-adjusted HR (95% CI) was 1.6 (1.1–2.5) for low BioE/high SHBG plus low 25(OH)D. Fracture risk for men with isolated low serum 25(OH)D, or those with low BioT with 25(OH)D > 50 nmol/L, did not differ from risk for men without low serum 25(OH)D or SH/SHBG abnormality. Significantly higher fracture risk was detected in the men with low BioE and/or high SHBG concurrent with a low 25(OH)D (adjusted HR 1.62; 95% CI: 1.05–2.51). The Panel notes that, in these older men, the fracture risk associated with baseline serum 25(OHD) concentration < 50 nmol/L (lowest quartile, mean 38.8 nmol/L) was observed only in the presence of low BioE or high SHBG, whereas 25(OH)D concentration < 50 nmol/L in isolation was not associated with fracture risk.

In a prospective cohort study, Rouzi et al. (2012) followed a cohort of 707 healthy Saudi post-menopausal women (mean age ± SD: 61.3 ± 7.2 years) for a mean ± SD of 5.2 ± 1.3 years. Their mean baseline serum 25(OH)D concentration was about 34 nmol/L. In multivariate logistic regression, besides physical activity score, age, hand-grip strength, BMD of total hip, past year history of falls, baseline serum 25(OH)D concentration and dietary calcium intake in the lowest quartiles were identified as independent predictors of risk of all osteoporosis-related fractures. For the lowest quartile (Q1) of serum 25(OH)D (≤ 17.9 nmol/L) vs higher values, the relative risk (RR) was 1.63 (95% CI: 1.06–2.51, p < 0.027) and for dietary calcium intake in Q1 (≤ 391 mg/day) vs higher values, RR was 1.66 (95% CI: 1.08–2.53, p < 0.020). The Panel notes that this study in post-menopausal women indicated an increase in the risk for osteoporosis-related fractures at baseline serum 25(OH)D concentration ≤ 17.9 nmol/L (lowest quartile).

In a pooled US cohort of 4,749 men and women aged 65 years and older from two surveys, Looker (2013) found that baseline serum 25(OH)D concentration was a significant linear predictor of risk of major osteoporotic fracture (hip, spine, radius and humerus) and significant quadratic predictor of hip fracture in the total sample and among those with less than 10 years of follow-up. It was not related to risk of either fracture type among those with 10 years of follow-up or more. After adjustments for potential confounders, fracture risk was significantly increased for serum 25(OH)D concentration < 30 nmol/L (major osteoporotic fracture RR: 2.09; 95% CI: 1.32–3.32; hip fracture RR: 2.63; 95% CI: 1.60–4.32), compared to serum 25(OH)D ≥ 30 nmol/L. Using other cut-off values, risk for either fracture outcome among those with serum 25(OH)D concentration between 30 and 49 nmol/L and 50 and 74 nmol/L did not differ from that seen in those with serum 25(OH)D ≥ 75 nmol/L, whereas the risk for either fracture was again significantly higher for those with serum 25(OH)D < 30 nmol/L. The Panel notes that this study in older subjects indicated an increase in the risk for fractures (major osteoporotic or hip only) at baseline serum 25(OH)D concentration < 30 nmol/L.

Using a stratified case-cohort design in 21,774 men and women (65–79 years) who attended four community-based health studies in Norway with a maximum follow-up of 10.7 years, Holvik et al. (2013) found an inverse association between baseline serum 25(OH)D concentration and risk of hip fracture. After adjustments for potential confounders, in the fully adjusted model (not adjusted for sex), only subjects with 25(OH)D concentration in the lowest quartile (< 42.2 nmol/L) had a 34% (95% CI: 5–70%) increased risk of hip fracture compared with the highest quartile (≥ 67.9 nmol/L). Investigating possible sex differences, after a first partial adjustment for age, sex, study centre and BMI, the association was statistically significant in men (HR 1.65; 95% CI: 1.04–2.61), but not in women, but the association was not statistically significant in either sexes in the fully adjusted model (including also month of blood sample). The Panel notes that, in this study in older subjects, an increased risk of hip fracture with baseline 25(OH)D concentration < 42 nmol/L (lowest quartile) was observed, when compared to 25(OH)D concentration ≥ 68 nmol/L (highest quartile).

In a population-based, prospective cohort study in Australia, Bleicher et al. (2014) followed 1,662 community-dwelling men (70–97 years) for a mean of 4.3 years (mean baseline serum 25(OH)D: about 56 nmol/L). In multivariate analyses,19 the risk of incident fractures was greatest in men with baseline 25(OH)D concentration in the lowest quintile (25(OH)D ≤ 36 nmol/L; mean 28.1 ± 6.6 nmol/L; HR: 3.5; 95% CI: 1.7–7.0) and in men in the highest quintile (25(OH)D > 72 nmol/L; HR: 2.7; 95% CI: 1.3–5.4), compared with men in the fourth quintile (25(OH)D ≥ 60 to ≤ 72 nmol/L). The difference in risk in quintiles 2 and 3 compared to 4 generally remained not statistically significant after additional adjustments20 or a sensitivity analysis. The Panel notes that this study in older men indicated an increased risk for fractures in men at baseline serum 25(OH)D concentration < 36 nmol/L and > 72 nmol/L (lowest and highest quintiles).

In a prospective study of 5,764 men and women, aged 66–96 years (either frail or healthy), based on a representative sample of the population of Reykjavik, Iceland, HRs of incident hip fractures were determined according to serum concentrations of 25(OH)D at baseline (Steingrimsdottir et al., 2014). Mean follow-up was 5.4 years. Compared with serum 25(OH)D of 50–75 nmol/L, HRs for hip fractures were 2.08 (95% CI 1.51–2.87) for serum 25(OH)D < 30 nmol/L in the fully adjusted model including physical activity. No difference in risk was associated with 30–50 nmol/L or ≥ 75 nmol/L in either model compared with the reference. This was also true when analysing men and women separately. The Panel notes that, in this study in older subjects, at baseline 25(OH)D concentration of < 30 nmol/L, the risk for hip fractures increased, whereas no difference in the risk was observed over the range above 30–75 nmol/L.

In a US prospective cohort study in 922 women during the menopausal transition and with an average follow-up of 9.5 years, Cauley et al. (2015) (Section 5.1.2.1.1) also determined if higher baseline 25(OH)D concentration is associated with lower fracture risk. The mean 25(OH)D concentration was 54.5 nmol/L; 43% of the women had 25(OH)D concentration < 50 nmol/L. There was no significant association between serum 25(OH)D and traumatic fractures. However, in multivariable adjusted hazards models, the HR for non-traumatic fractures was 0.72 (95% CI: 0.54–0.95) for each 25 nmol/L increase in 25(OH)D, and was 0.54 (95% CI: 0.32–0.89) when comparing women whose 25(OH)D concentration was ≥ 50 vs < 50 nmol/L. The Panel notes that, in this study, serum 25(OH)D concentration < 50 nmol/L was associated with an increased risk for non-traumatic fracture in mid-life women.

Conclusions on fracture risk in adults

Among the 15 recent prospective observational studies identified, most of which were in older free-living adults, the Panel notes the heterogeneity of observational study designs, populations and fracture sites investigated and considers that the relationship of serum 25(OH)D concentration and fracture risk may be confounded by a variety of factors (see Section 5.1.2.1.1). Furthermore, observational studies mostly used single measurements of 25(OH)D concentration, thus possible long-term changes in 25(OH)D concentration were not considered in the analyses of the relationship with fracture risk.

An increased risk of fractures was seen at baseline 25(OH)D concentrations < 18 nmol/L (Rouzi et al., 2012) (lowest quartile), < 30 nmol/L (Looker, 2013; Steingrimsdottir et al., 2014), < 36 nmol/L (Bleicher et al., 2014) (lowest quintile), < 38 nmol/L (Robinson-Cohen et al., 2011), < 42 nmol/L (Holvik et al., 2013) (lowest quartile), < 50 nmol/L ((Cauley et al., 2015); lowest quartile in Cauley et al. (2010), lowest quartile and only in case of low sex steroid concentrations for Barrett-Connor et al. (2012)), and < 71 nmol/L (Nakamura et al., 2011) (quartiles Q1–Q3). One study observed a significant negative relationship between PTH concentration and fracture risk at serum 25(OH)D concentrations < 50–80 nmol/L (Rejnmark et al., 2011). An increased fracture risk was also reported at 25(OH)D concentrations > 72 nmol/L (Bleicher et al., 2014) (highest quintile), > 50 nmol/L in black women and > 75 nmol/L in Asian (non-statistically significant) women but a lower fracture risk at 25(OH)D < 75 nmol/L in white women (statistically significant) (Cauley et al., 2011). However, three studies found no difference in fracture risk between baseline serum 25(OH)D concentrations in the lowest quartile (< 45 nmol/L, (Barbour et al., 2012); < 50 nmol/L (Bolland et al., 2010); < 63 nmol/L, (Chan et al., 2011)) and higher concentrations.

The Panel notes that 9 out of 15 observational studies reported an increased risk for fractures that was associated with baseline 25(OH)D concentrations between < 18 nmol/L and < 50 nmol/L in free-living adult populations (Cauley et al., 2010, 2015; Robinson-Cohen et al., 2011; Barrett-Connor et al., 2012; Rouzi et al., 2012; Holvik et al., 2013; Looker, 2013; Bleicher et al., 2014; Steingrimsdottir et al., 2014). One study observed a significant negative relationship between PTH concentration and fracture risk at serum 25(OH)D concentrations < 80 nmol/L (Rejnmark et al., 2011) and, in one study in Japanese women (with low calcium intake), an increased fracture risk was reported at 25(OH)D concentration < 71 nmol/L (Nakamura et al., 2011).

In contrast, an increased fracture risk was observed at ≥ 50 to ≥ 75 nmol/L in two studies: in Bleicher et al. (2014), as well as in Cauley et al. (2011) only in black (significant result) and Asian (non-significant result) women, respectively. This was not observed in other studies: in Cauley et al. (2011) in white women, nor in Chan et al. (2011), Barbour et al. (2012) or Looker (2013).

The Panel notes the conclusions by IOM (2011) on a wide variation in serum 25(OH)D concentration associated with an increased fracture risk. Taking into account also the observational studies published thereafter (mainly on older adults), the Panel considers that, overall, the majority of studies indicate an increased fracture risk associated with 25(OH)D concentrations of < 18 nmol/L to < 50 nmol/L in free-living adults.

5.1.2.1.4 Muscle strength/function and physical performance

IOM (2011) (Section 4 and Appendix B) considered physical performance and falls as independent health outcomes, but because of the joint consideration of these outcomes in the literature, the available evidence was considered together. IOM (2011) reported some support, mainly from observational studies, for an association between 25(OH)D concentrations and physical performance, but concluded that high-quality observational evidence from larger cohort studies was lacking.

Lamberg-Allardt et al. (2013) identified two systematic reviews with meta-analyses of RCTs on vitamin D and muscle strength in older subjects (Muir and Montero-Odasso, 2011; Stockton et al., 2011). Based on a meta-analysis of 17 RCTs (n = 5,072, mean age 60 years in most studies), Stockton et al. (2011) concluded that vitamin D supplementation does not have an effect on muscle strength in adults with mean baseline serum 25(OH)D concentrations ≥ 25 nmol/L, and that two RCTs (in patients) demonstrate an increase in hip muscle strength in adults with serum 25(OH)D concentrations < 25 nmol/L. The systematic review on 13 RCTs (n = 2,268) by Muir and Montero-Odasso (2011) concluded that vitamin D doses of 20–25 μg/day showed beneficial effects on balance and muscle strength in older adults (≥ 60 years of age). Mean baseline serum 25(OH)D concentrations were about 25–65 nmol/L in 12 RCTs that provided the information (mean baseline of 25–50 nmol/L in 10 of these RCTs). The Panel notes that only three references among the studies considered in these two systematic reviews were published in 2010 or afterwards, and that seven RCTs were in common in both systematic reviews.

Newberry et al. (2014) identified two new RCTs in older adults that examined the effects of one year of vitamin D supplementation with calcium on muscle strength or function (Pfeifer et al., 2009; Zhu et al., 2010). Newberry et al. (2014) also identified five prospective cohort studies on the association between serum 25(OH)D concentrations and muscle strength, muscle function or physical performance (Dam et al., 2009; Scott et al., 2010; Michael et al., 2011; Houston et al., 2012; Menant et al., 2012). Newberry et al. (2014) concluded that the associations between serum 25(OH)D concentrations and muscle strength, muscle function or physical performance in post-menopausal women or older men were inconsistent.

SACN (2016) considered three systematic reviews with meta-analyses of RCTs (two already mentioned above (Muir and Montero-Odasso, 2011; Stockton et al., 2011) and another one (Beaudart et al., 2014)21 on 30 RCTs (n = 5,615). These systematic reviews reported a beneficial effect of vitamin D supplementation on muscle strength and function in adults aged > 50 years with mean baseline serum 25(OH)D concentrations of 24–66 nmol/L (Muir and Montero-Odasso, 2011), < 30 nmol/L (Beaudart et al., 2014), and < 25 nmol/L (institutionalised older adults (Stockton et al., 2011)). The Panel notes that 14 RCTs out of the 30 RCTs included in (Beaudart et al., 2014) were published in 2010 or afterwards,22 and 8 or 11 references were in common with the systematic review by Muir and Montero-Odasso (2011) or by Stockton et al. (2011), respectively. SACN (2016) identified three subsequent RCTs (Lips et al., 2010; Knutsen et al., 2014; Pirotta et al., 2015) and seven cohort studies (Bolland et al., 2010; Scott et al., 2010; Houston et al., 2011, 2012; Michael et al., 2011; Chan et al., 2012; Menant et al., 2012), which provided mixed results, and also noted that, in most of the cohort studies, cut-offs were predefined.

The Panel considered pertinent primary studies from 2010 onwards mostly on healthy adults and, when excluding studies in populations during a resistance training intervention, retrieved 15 intervention and prospective observational studies, reporting on muscle strength or function, physical performance or related outcomes (e.g. postural stability, muscle power, mobility), in relation to 25(OH)D concentrations. In the following section, the nine intervention studies and then the six prospective observational studies are described individually. The results are then summarised, and an overall conclusion on muscle strength/function and physical performance is provided.

RCTs with vitamin D supplementation

In a 16-week double-blind multicentre RCT in North America and Europe, Lips et al. (2010) studied the effects of a dose of 210 μg vitamin D₃ per week (~ 30 μg/day) or a placebo on postural stability, measured as postural body sway, and physical performance, measured as short physical performance battery (SPPB),23 in 246 older subjects (age 70 years and older). Baseline serum 25(OH)D concentrations were between 15 and 50 nmol/L. Mean serum 25(OH)D concentration increased significantly from 35 to 65 nmol/L (p < 0.001) in supplemented subjects, with no change in the placebo group. No differences in postural stability or physical performance were observed between groups at the end of the study. In a post-hoc analysis of a subgroup of subjects with elevated sway at baseline, supplementation with vitamin D₃ significantly reduced sway. The Panel notes that this study in older subjects with weekly vitamin D₃ supplementation, which increased their mean serum 25(OH)D concentration from 35 to 65 nmol/L, found no effect on postural stability or physical performance compared with placebo. The Panel also notes that the study found an increased postural stability (i.e. significantly reduced sway) in a subgroup of subjects with elevated body sway at baseline.

In a 6-month double-blind RCT in the Netherlands, Janssen et al. (2010) compared the effects of a daily supplementation of 10 μg vitamin D₃ and 500 mg calcium with a placebo + 500 mg calcium supplementation only, on muscle strength (knee extension or handgrip strength), power (leg extension power) and mobility (Timed Up And Go (TUAG) test and Modified Cooper test24) in 70 female geriatric outpatients. Most participants lived in residential homes, all were above 65 years of age with baseline serum 25(OH)D concentrations between 20 and 50 nmol/L (mean baseline of 33–34 nmol/L among groups). At 6 months, a significant difference in mean serum 25(OH)D (77.2 vs 41.6 nmol/L, p < 0.001) and 1,25(OH)₂D concentrations (94.1 vs 67.5 pmol/L, p < 0.001) was found between the two groups, but no differences in muscle strength, power or mobility. The Panel notes that, in this study, older subjects supplemented daily with vitamin D₃ and calcium for 6 months, compared with calcium alone, increased their mean serum 25(OH)D from 33 to 77 nmol/L compared with increases from 34 to 42 nmol/L in the placebo + calcium group, and that no effect on muscle strength, power or mobility was observed.

In a 1-year population-based double-blind RCT in Australia, Zhu et al. (2010) assessed the effects of a daily 25 μg vitamin D₂ supplement or placebo (both groups receiving 1 g calcium/day) on muscle strength in different muscle groups and mobility using the TUAG test in 302 older community-dwelling women aged 70–90 years. Mean baseline serum 25(OH)D was 44 ± 10.5 nmol/L (with 66% of subjects with 25(OH)D concentration lower than 50 nmol/L). In the vitamin D and calcium group after 1 year, 25(OH)D concentration increased to 60 ± 14 nmol/L (with 80% of subjects achieving a serum 25(OH)D concentration higher than 50 nmol/L). For hip extensor and adductor strength and TUAG, but not for other muscle groups, a significant interaction between treatment group and baseline values of 25(OH)D was noted. Only in subjects in the lowest tertile of baseline hip extensor and adductor strength and TUAG test, muscle strength and TUAG test improved more with vitamin D and calcium supplementation compared with calcium supplementation alone. Baseline 25(OH)D concentration did not influence subject's response to supplementation with regard to muscle strength and mobility. The Panel notes that this study in older women supplemented daily with vitamin D₂ together with calcium for 12 months increased mean serum 25(OH)D concentration from 44 to 60 nmol/L, compared with calcium alone, and that increased muscle strength and mobility were found only in those who were the weakest and slowest at baseline.

In a 6-month double-blind, randomised exploratory clinical trial in the USA, Lagari et al. (2013) investigated the effects of daily 10 or 50 μg vitamin D₃ supplementation on physical performance and muscle strength, in 86 community-dwelling subjects aged 65–95 years with a mean baseline serum 25(OH)D concentration of 82.5 nmol/L. Physical performance was assessed as a four-metre walk speed test to calculate gait speed, timed sit-to-stand test or chair stand test, single-leg balance test and gallon-jug test, and muscle strength was measured as handgrip test. A mean decrease in serum 25(OH)D concentration of 3 nmol/L in men (n = 6) and 8.5 nmol/L in women (n = 25) was observed in the 10 μg/day supplement group and a mean increase was observed in the 50 μg/day supplement group of 16 nmol/L in men (n = 9) and 13 nmol/L in women (n = 46). Overall, no significant changes in physical performance or muscle strength were found at the end of the intervention period. However, subjects with the slowest gait speed at baseline improved their ability to do chair-stand tests after vitamin D supplementation, after adjustments for potential confounders. The Panel notes that, in this study in older subjects, two daily doses of vitamin D₃ supplementation for 6 months decreased (−3 to −8.5 nmol/L) or increased serum 25(OH)D concentrations (+13 to +16 nmol/L) from a mean baseline of 82.5 nmol/L, and that no effect of dose on physical performance or muscle strength was measured. The study showed that subjects with the slowest gait speed at baseline showed an improvement in one of the physical performance tests.

In a 12-week RCT in the UK in 25 young athletes (mean age 21 years) receiving either placebo, 500 μg or 1,000 μg/week vitamin D₃ (~ 71 μg/day and 142 μg/day), Close et al. (2013b) measured serum 25(OH)D concentration and muscle function (bench press and leg press and vertical jump height) before supplementation and at 6 and at 12 weeks post-supplementation. Baseline mean serum 25(OH)D concentration was 51 ± 24 nmol/L, with 57% of subjects below 50 nmol/L. Following 6 and 12 weeks supplementation, serum 25(OH)D concentration increased above 50 nmol/L in all participants (mean in each group: about 85–90 nmol/L (values read on figure)). In contrast, 25(OH)D concentration in the placebo group decreased at 6 and 12 weeks to 37 ± 18 and 41 ± 22 nmol/L, respectively. None of the muscle function parameters in these young athletes was significantly affected by an increase of serum 25(OH)D concentration. The Panel notes that, in younger subjects, weekly vitamin D₃ supplementation for 12 weeks increased their serum 25(OH)D concentration above 50 nmol/L, and that this study found no effect on muscle function compared with placebo.

In a parallel group double-blind RCT by Wood et al. (2014) (Sections 5.1.2.1.5, 5.2.3 and 5.2.4), 'healthy' post-menopausal women from North East Scotland aged 60–70 years, were assigned to daily vitamin D₃ of 10 μg (n = 102), of 25 μg (n = 101) or matching placebo (n = 102) for 1 year. Grip strength (primary outcome), diet, physical activity and UV-B irradiation exposure were measured bimonthly, as were serum 25(OH)D, phosphate and calcium and concentrations. Mean (SD) serum 25(OH)D concentrations at baseline were 34.3 (14.7) nmol/L, 33.9 (14.3) nmol/L and 32.4 (16.3) nmol/L in normal weight (BMI < 25 kg/m²; n = 113), overweight (BMI 25–25.99 kg/m²; n = 139) and obese (BMI ≥ 30 kg/m²; n = 53) subjects, respectively. After 1 year of treatment with 10 and 25 μg of vitamin D, serum 25(OH)D concentration increased among the various BMI groups by 31–33 μmol/L and 39–48 nmol/L, respectively. In contrast, the change in 25(OH)D in the placebo groups was between − 1.7 to − 6.6 μmol/L. Supplementation had no effect on grip strength. The Panel notes that, in this study, two different daily doses of vitamin D₃ supplementation for 1 year increased mean serum 25(OH)D concentration, but had no effect on grip strength compared to placebo.

In a 16-week randomised, double-blind, placebo-controlled trial in Norway, Knutsen et al. (2014) compared the effects of a daily vitamin D₃ supplementation (10 or 25 μg vitamin D₃) or placebo on muscle power and strength measured as jump height and handgrip strength and chair-rising differences between pre- and post-intervention in adults from ethnic minority groups (n = 215) with a mean age of 37 years (range 18–50 years). Mean serum 25(OH)D₃ concentration increased from 27 to 52 nmol/L and from 27 to 43 nmol/L in the groups receiving 25 and 10 μg/day, respectively, with no changes in the placebo group. Vitamin D supplementation had no significant effect on muscle power or strength. The Panel notes that this 16-week study in younger adults from minority ethnic groups with two daily supplemental doses of vitamin D₃ increased mean 25(OH)D concentration from 27 to 52 or 43 nmol/L with no significant effect on muscle power or muscle strength compared with placebo.

In a 10-week RCT in Australia, Pirotta et al. (2015) investigated the effects of a daily supplement (50 μg vitamin D₃ or a placebo) in 26 older adults (> 60 years) with baseline 25(OH)D concentrations between 25 and 60 nmol/L on neuroplasticity as the primary outcome and muscle power and function (mobility) measured as stair climbing power, gait (TUAG), dynamic balance (four square step test) as the secondary outcome. Mean serum 25(OH)D concentration increased from 46 to 81 nmol/L in the vitamin D supplemented group with no changes in the placebo group. No significantly different changes in any of the outcome measures were observed between the vitamin D supplemented and placebo groups at the end of the intervention period. The Panel notes that this was a relatively short intervention study and that it showed that daily vitamin D supplementation increased mean serum 25(OH)D concentration from 46 to 81 nmol/L with no effect on muscle power or function in older adults compared with placebo.

In a 2-year double-blind RCT in 409 home-dwelling women (70–80 years) in Finland, Uusi-Rasi et al. (2015) (Sections 5.1.2.1.1, 5.1.2.1.5, 5.2.3 and 5.2.4) also investigated the effect of daily vitamin D₃ supplementation (20 μg) with or without exercise (compared with placebo groups, with or without exercise), on measures of physical functioning (secondary analysis). These measures were physical performance (measured by SPPB), mobility (TUAG test), dynamic balance (backward walking) and muscle strength (leg extensor strength). These subjects were at risk of falling (having fallen at least once during the previous 12 months), and were not performing vigorous exercise more than 2 h per week. Mean serum 25(OH)D concentrations were not significantly different among groups at baseline, between 65.5 and 69.5 nmol/L, and, after 2 years, they remained almost unchanged in the two placebo groups (around 68.7 nmol/L), while, in the vitamin D groups, they increased up to 92.5 nmol/L. Vitamin D supplementation alone did not improve any measures of physical functioning (which were improved by exercise). The Panel notes that, in this study in women at risk of falling, daily vitamin D supplementation alone increased mean serum 25(OH)D concentration from about 66 to 92.5 nmol/L with no improvement in physical functioning.

Prospective observational studies

In a cohort of 686 community-dwelling older adults (mean age 62 ± 7 years, 49% women) in Australia, Scott et al. (2010) investigated associations between serum 25(OH)D concentration and leg muscle strength and leg muscle quality (LMQ)25 at baseline and at a mean follow-up of 2.6 ± 0.4 years. At baseline, 297 subjects had serum 25(OH)D concentration ≤ 50 nmol/L (mean ± SD of 37.1 ± 8.4 nmol/L), and 389 had serum 25(OH)D > 50 nmol/L (mean ± SD of 67.8 ± 13.4 nmol/L). After adjustments for potential confounders (including season of analysis), baseline 25(OH)D concentration was positively associated with the change in leg muscle strength and LMQ over 2.6 years (p = 0.027 and 0.003, respectively). The Panel notes that, in this study in older adults in which about 43% had baseline serum 25(OH)D below 50 nmol/L, baseline 25(OH)D concentration was positively associated with the change in leg muscle strength and LMQ.

Bolland et al. (2010) (Section 5.1.2.1.1, 5.1.2.1.3 and 5.1.2.1.5) examined the association between baseline serum 25(OH)D concentration and multiple health outcomes in 1,471 community-dwelling women (mean age 74 years) who took part in a 5-year calcium supplementation study in Australia. Fifty per cent of women had a seasonally adjusted 25(OH)D concentration < 50 nmol/L. After adjustments for potential confounders (including treatment allocation to calcium or placebo), women with a seasonally adjusted baseline 25(OH)D concentration < 50 nmol/L and those with 25(OH)D concentration ≥ 50 nmol/L did not show any difference in change (decline) in grip strength. The Panel notes that this study of community-dwelling older adults showed no difference in change in grip strength in women with a seasonally adjusted baseline 25(OH)D concentration < 50 nmol/L compared with those with 25(OH)D concentration ≥ 50 nmol/L, over a 5-year period.

In a cohort of 534 US post-menopausal women (mean age: 70.3 ± 3.9 years, mainly Caucasian), Michael et al. (2011) evaluated the association between baseline serum 25(OH)D concentration (48.2 ± 21.4 nmol/L) and a physical summary score at baseline, at 1, 3 and 6 years. The physical summary score was derived from data on timed walk test, chair-stand test and grip strength. In the 6 years of follow-up, participants with baseline serum 25(OH)D concentration ≥ 75 nmol/L (but not those with 25(OH)D of 25–49 and 50–74 nmol/L) had significantly higher scores for physical performance compared with the reference category (< 25 nmol/L) after adjustments for potential confounders (p < 0.001). Physical performance declined over the follow-up period as a result of ageing, but higher baseline serum 25(OH)D concentration was not associated with a reduction in the decline in physical performance over the 6-year period. The Panel notes that this study showed that higher baseline serum 25(OH)D concentration (≥ 75 nmol/L) in older women was associated with higher physical performance at follow-up compared with baseline concentration < 25 nmol/L, but was not associated with the age-related decline in physical performance over a 6-year period.

In community-dwelling men and women aged 77–100 years in four different US settings, Houston et al. (2011) examined the association between baseline serum 25(OH)D concentration and mobility disability (difficulty walking half a mile or up 10 steps) and activities of daily living (ADL) disability measured at baseline and every 6 months over 3 years of follow-up (longitudinal analysis). Almost one-third (31%) of participants had serum 25(OH)D concentration < 50 nmol/L at baseline. After adjustments for potential confounders, in participants free of mobility disability at baseline, participants with baseline serum 25(OH)D concentration < 50 nmol/L (but not participants with serum 25(OH)D of 50–74 nmol/L) were at greater risk of incident mobility disability over 3 years of follow-up (HR: 1.56; 95% CI: 1.06–2.30), compared with those with serum 25(OH)D concentration ≥ 75 nmol/L. In participants free of ADL disability at baseline, there was no association between baseline serum 25(OH)D concentration and risk of ADL disability. The Panel notes that, in this study in older community-dwelling adults, participants with baseline serum 25(OH)D concentration < 50 nmol/L had a greater risk of incident mobility disability (but not of ADL disability) after 3 years of follow-up compared with those with serum 25(OH)D ≥ 75 nmol/L.

In a cohort of 2,641 men and women (age 71–80 years), 38% African American, in the USA, Houston et al. (2012) investigated associations between serum 25(OH)D concentration measured at baseline and physical performance, measured as SPPB and a second physical performance battery, gait speed (20-m or 400-m), and muscle strength (knee extensor strength and grip strength), measured at baseline and at 2 and 4 years follow-up. After full adjustments for potential confounders, longitudinal associations between baseline 25(OH)D concentration and physical performance at 4-year follow-up showed that participants with serum 25(OH)D < 50 nmol/L (but not those with serum 25(OH)D of 50–74 nmol/L) had poorer physical performance than participants with 25(OH)D ≥ 75 nmol/L (p < 0.01 for both battery scores) and lower 400-m gait speed (p < 0.001). Baseline serum 25(OH)D was not associated with muscle strength at the 4-year follow-up. Physical performance and gait speed declined over the 4 years of follow-up (p < 0.0001), and, except for SPPB, the rate of decline was not associated with baseline 25(OH)D concentration. The Panel notes that this study in older subjects showed a poorer physical performance at 4 years (but not muscle strength) in subjects with baseline serum 25(OH)D concentration < 50 nmol/L compared with ≥ 75 nmol/L, but that serum 25(OH)D concentration at baseline was not related to the age-related decline in physical performance and strength over a 4-year follow-up.

In a longitudinal analysis of a prospective cohort study in China of community-dwelling men (n = 714; age > 65 years), Chan et al. (2012) analysed the association between baseline serum 25(OH)D concentration and 4-year physical performance measures (including grip strength, 6-m walking speed, step length in a 6-m walk, time to complete five chair stands). Baseline mean ± SD serum 25(OH)D concentration was 77.9 ± 20.5 nmol/L with 94% of participants having a concentration of 50 nmol/L or greater. After adjustment for potential confounding factors, serum 25(OH)D concentration was not associated with baseline or 4-year absolute change in physical performance measures. The Panel notes that this study in older community-dwelling men with relative high baseline serum 25(OH)D concentration showed no association with physical performance after a 4-year period.

Conclusions on m uscle strength/function and physical performance in adults

The Panel notes the heterogeneity in the design of the nine RCTs with respect to age profile of subjects, dose and length of administration of vitamin D with or without calcium, and measures of muscle strength and physical performance or related outcomes. The Panel notes that seven RCTs were carried out in older not-institutionalised subjects (Janssen et al., 2010; Lips et al., 2010; Zhu et al., 2010; Lagari et al., 2013; Wood et al., 2014; Pirotta et al., 2015; Uusi-Rasi et al., 2015).

The Panel notes that, in the nine RCTs with vitamin D supplementation (with or without calcium) between 10 weeks and 2 years, mean serum 25(OH)D concentrations increased from 27 nmol/L (Knutsen et al., 2014), 33 nmol/L (Janssen et al., 2010), about 32–34 nmol/L (Wood et al., 2014), 35 nmol/L (Lips et al., 2010), 44 nmol/L (Zhu et al., 2010), 46 nmol/L (Pirotta et al., 2015), 51 nmol/L (Close et al., 2013b), about 66 nmol/L (Uusi-Rasi et al., 2015) or 82.5 nmol/L (Lagari et al., 2013), up to 52 nmol/L, 77 nmol/L, about 82 nmol/L, 65 nmol/L, 60 nmol/L, 81 nmol/L, about 90 nmol/L, about 92.5 nmol/L or about 98 nmol/L, respectively. These RCTs showed that increasing mean serum 25(OH)D concentrations from these baseline to final values by vitamin D supplementation did not result in a change in measures of physical performance or muscle strength/function.

The Panel notes that all six prospective observational studies identified on the association between baseline serum 25(OH)D concentration and muscle strength/physical performance were on older subjects, but otherwise were heterogeneous with respect to design, and that the studies may be confounded by a variety of factors (Sections 5.1.2.1.1 and 5.1.2.1.3). Furthermore, as for other health outcomes (Sections 5.1.2.1.1 and 5.1.2.1.3), observational studies used single measurements of 25(OH)D concentration, thus possible long-term changes in 25(OH)D concentration were not considered in the analyses of the relationship with muscle strength/physical performance.

In one study in older adults in which about 43% had baseline serum 25(OH)D below 50 nmol/L, baseline 25(OH)D concentration was positively associated with the change in leg muscle strength and LMQ (Scott et al., 2010). Three other observational studies (Houston et al., 2011, 2012; Michael et al., 2011) used predefined cut-off concentration for serum 25(OH)D, of < 25 nmol/L (versus 25–49, 50–74 and ≥ 75 nmol/L) (Michael et al., 2011), or > 75 nmol/L (versus < 50 or 50–74 nmol/L) (Houston et al., 2011, 2012). Among these three studies, two studies showed a higher risk of mobility disability as well as poorer physical performance in men and women with baseline serum 25(OH)D concentrations below 50 nmol/L (versus ≥ 75 nmol/L) (Houston et al., 2011, 2012). A third study, in older women, showed a better physical performance at 6-year follow-up with baseline serum 25(OH)D concentrations ≥ 75 nmol/L (versus < 25 nmol/L) (Michael et al., 2011). In contrast, one study showed no difference in change in muscle strength (grip strength) in women with a seasonally adjusted baseline 25(OH)D concentration < 50 nmol/L (predefined cut-off) compared with those with 25(OH)D concentrations ≥ 50 nmol/L (Bolland et al., 2010). Finally, one study showed no association between serum 25(OH)D (mean baseline: 78–94 nmol/L) and measures of physical performance (Chan et al., 2012). The Panel notes that the observational studies were inconsistent in their findings.

In its conclusion, the Panel took into account the conclusions by IOM (2011) on some (mainly observational) evidence supporting an association between serum 25(OH)D concentrations and physical performance and on the lack of large high-quality observational evidence, the conclusions of Lamberg-Allardt et al. (2013), Newberry et al. (2014) and SACN (2016). The Panel also took into account the identified studies published thereafter, and notes that the evidence is inconsistent. The Panel considers that, overall, the recent RCTs, all undertaken in populations with mean baseline serum 25(OH)D concentration of 27 nmol/L or higher, show no support for an association between serum 25(OH)D concentration and physical performance in healthy older adults. Four of the six new prospective observational studies used predefined cut-off values for serum 25(OH)D concentration. The Panel considers that four out of six observational studies reported a positive association between baseline serum 25(OH)D and better muscle strength/quality, lower risk of mobility disability or of poorer physical performance at follow-up. Overall, from the available evidence, the Panel considers that no target value for serum 25(OH)D concentration with regard to muscle strength/function and physical performance can be derived.

5.1.2.1.5 Risk of falls and falling

A fall is defined as 'the unintentional coming to rest on the ground, floor, or other lower level' and the number of falls in a population subgroup over a period of time can be recorded and results expressed as, e.g. the number of falls per person per observation time (incidence), the total number of falls or the number of subjects falling at least once (termed fallers) (EFSA NDA Panel, 2011)).

IOM ( 2011 ) (Section 4 and Appendix B) concluded that the greater part of RCTs found no effects of vitamin D with or without calcium on reduction in the risk for falls. IOM (2011) also concluded that the observational studies (mostly cross-sectional) suggested an association between a higher serum 25(OH)D concentration and a reduced risk of falls in older adults.

Lamberg-Allardt et al. (2013) based their conclusions on seven systematic reviews (Cranney et al., 2007; Chung et al., 2009; Kalyani et al., 2010; Michael et al., 2010; Murad et al., 2011; Cameron et al., 2012; Gillespie et al., 2012). Lamberg-Allardt et al. (2013) noted that the systematic reviews included many of the same studies, with some variation due to different inclusion and exclusion criteria and timeframe, and that the definition of 'falls' and 'falling' varied among trials. Lamberg-Allardt et al. (2013) concluded that there is probable evidence that supplementation with vitamin D in combination with calcium is effective in preventing falls in older adults, especially in those with 'low' baseline serum 25(OH)D concentration either community-dwelling or in nursing care facilities. The threshold for a 25(OH)D concentration below which the risk for falls or falling was increased was unclear.

Newberry et al. (2014) identified two RCTs, already cited in the IOM report, and that examined the effect of supplementation with vitamin D and calcium on the risk of falls/falling among older adults (Prince et al., 2008; Pfeifer et al., 2009), as well as one prospective cohort study (Menant et al., 2012) on serum 25(OH)D concentration and the risk of falls. Newberry et al. (2014) concluded that an association was seen between lower serum 25(OH)D concentration and increased risk of falls.

SACN (2016) considered five systematic reviews and meta-analyses (Kalyani et al., 2010; Murad et al., 2011; Cameron et al., 2012; Gillespie et al., 2012; Bolland et al., 2014a), two RCTs of which one (Sanders et al., 2010) was considered by IOM (2011), and the other one (Bischoff-Ferrari et al., 2016) used supplementation given monthly which did not correspond to the inclusion criteria defined by the Panel for its literature search (Section 5.1). SACN (2016) also considered one cohort study (Menant et al., 2012), and two genetic studies (Onder et al., 2008; Barr et al., 2010).

SACN (2016) concluded that the evidence on vitamin D and falls is mixed but, on balance, that the evidence is suggestive of beneficial effects of vitamin D supplementation in reducing fall risk in adults > 50 years with mean baseline serum 25(OH)D concentrations over a broad range of values (23–59, 24–28, 24–55, 23–82 nmol/L according to the systematic reviews considered).

In addition to the RCTs by Wood et al. (2014) and Uusi-Rasi et al. (2015) (Sections 5.1.2.1.1, 5.1.2.1.4, 5.2.3 and 5.2.4) and the observational study by Bolland et al. (2010) (Sections 5.1.2.1.1, 5.1.2.1.3 and 5.1.2.1.4), the Panel identified one prospective observational study in free-living older adults published after the IOM report, that is described hereafter and followed by an overall conclusion on risk of falls and falling.

RCTs with vitamin D supplementation

In the double-blind RCT in 'healthy' post-menopausal women from Scotland (60–70 years, BMI 18–45 kg/m²) assigned to daily vitamin D₃ of 10 μg (n = 102), 25 μg (n = 101) or matching placebo (n = 102) for 1 year (mean baseline serum 25(OH)D: about 32–34 nmol/L) (Sections 5.1.2.1.4, 5.2.3 and 5.2.4), Wood et al. (2014) also measured falls bimonthly (secondary outcome, self-reported at study visit) among the various BMI groups. The Panel notes that, in this study, two different daily doses of vitamin D₃ supplementation for 1 year increased mean serum 25(OH)D concentration, but had no effect on the number of 'ever fallen' falls compared to placebo.

In the 2-year double-blind RCT in 409 home-dwelling women in Finland (70–80 years, mean baseline serum 25(OH)D concentrations among groups between 65.5 and 69.5 nmol/L) who were at risk of falling, Uusi-Rasi et al. (2015) (Sections 5.1.2.1.1, 5.1.2.1.4, 5.2.3, and 5.2.4) also investigated the effect of daily vitamin D₃ supplementation (20 μg) with or without exercise on self registered falls (primary outcome) obtained from prospective fall diaries returned monthly via mail as primary outcome. There was no interaction between vitamin D and exercise. Fall incidence rate ratios calculated as the total number of falls divided by the time over which falls were monitored (100 persons/year), and HRs for fallers were not different between the two vitamin D and the two placebo groups. In a secondary analysis, HRs for injured fallers (faller requiring medical care) were significantly lower in both exercise groups compared with placebo without exercise, and the HR in the vitamin D group without exercise was not different from those in the placebo without exercise. The Panel notes that this study in home-dwelling older women found that daily vitamin D supplementation for 2 years increased mean serum 25(OH)D concentration, but had no effect on the rate or risk of falls or injurious falls.

Prospective observational studies

Bolland et al. (2010) examined the association between baseline serum 25(OH)D concentration and multiple health outcomes (Sections 5.1.2.1.1, 5.1.2.1.3 and 5.1.2.1.4) in 1,471 community-dwelling women (mean age 74 years) who took part in a 5-year calcium supplementation study in Australia. Fifty per cent of women had a seasonally adjusted 25(OH)D concentration < 50 nmol/L. After adjustments for potential confounders (including treatment allocation to calcium or placebo), women with a seasonally adjusted baseline 25(OH)D concentration < 50 nmol/L were not at increased risk of falls (monitored with a diary), compared with those with 25(OH)D concentration ≥ 50 nmol/L. The Panel notes that this study of community-dwelling older women with a seasonally adjusted 25(OH)D concentration < 50 nmol/L compared with those with 25(OH)D concentration ≥ 50 nmol/L showed no increased risk of falls over a 5-year period.

In a cohort of 463 older community-dwelling men and women (54%) (age 70–90 years) in Australia, Menant et al. (2012) studied the relationship between baseline serum 25(OH)D concentration and falls monitored with monthly diaries and assessed at 12-months follow-up. At baseline, 21% of men and 44% of women had serum 25(OH)D concentration ≤ 50 nmol/L. After adjustments for potential confounders, baseline serum 25(OH)D concentration < 50 nmol/L (predefined cut-off) was associated with an increased rate of falls in men (incident rate ratio: 1.93; 95% CI: 1.19–3.15, p = 0.008), but not in women. The Panel notes that this study in older subjects showed that serum 25(OH)D concentration < 50 nmol/L was associated with increased rate of falls in men only.

Conclusions on risk of falls and falling in adults

The Panel considered two RCTs published after the IOM report, which showed that mean serum 25(OH)D concentrations increased after vitamin D₃ supplementation for 1–2 years, while this supplementation had no effect on the number of 'ever fallen' falls compared to placebo in one study, or on the rate or risk of falls or injurious falls in the other. The Panel considered two prospective observational study published after the IOM report, with inconsistent results. One study of community-dwelling older women with a seasonally adjusted 25(OH)D concentration < 50 nmol/L compared with those with 25(OH)D concentration ≥ 50 nmol/L showed no increased risk of falls over a 5-year period (Bolland et al., 2010). The other study in older subjects showed that serum 25(OH)D concentration < 50 nmol/L was associated with increased rate of falls in men only (Menant et al., 2012). Furthermore, as for other health outcomes (Sections 5.1.2.1.1, 5.1.2.1.3 and 5.1.2.1.4), observational studies used single measurements of 25(OH)D concentration, thus possible long-term changes in 25(OH)D concentration were not considered in the analyses of the relationship with rate/risk of falls.

The Panel considered the conclusion by IOM (2011), by SACN (2016), Newberry et al. (2014), Lamberg-Allardt et al. (2013), that took several systematic reviews (undertaken with different inclusion criteria) into account. The Panel notes that the evidence on serum 25(OH)D is inconsistent, but overall, is suggestive of beneficial effects of vitamin D in reduction of the risk of falling in older adults over a broad range of mean baseline serum 25(OH)D concentrations (23–82 nmol/L according to the systematic reviews considered in previous reports). From the available evidence, the Panel concludes that no target value for serum 25(OH)D concentration with regard to the risk of falls or falling can be derived.

5.1.2.1.6 Calcium absorption

Regarding the physiological role of 1,25(OH)₂D in the active transport regulation of calcium absorption in the intestine (Section 2.2.1) (EFSA NDA Panel, 2015b), the Panel considered it pertinent to review the possible relationship between 25(OH)D concentration and calcium absorption to try to identify a possible threshold value for this relationship. Calcium absorption is usually measured as fractional calcium absorption for which the dual calcium isotope technique is regarded as the gold standard (Heaney, 2000; IOM, 2011), whereas single isotope methods, which are considered more convenient to use, have also been developed (Lee et al., 2011).

IOM ( 2011 ) (Section 4 and Appendix B) considered both RCTs and cross-sectional studies in relation to vitamin D status and calcium absorption and concluded that fractional calcium absorption reaches a maximum at serum 25(OH)D concentrations between 30 and 50 nmol/L in adults, 'with no clear evidence of further benefit above 50 nmol/L'. The Panel notes that the IOM included the study by Need et al. (2008) in patients attending osteoporotic clinics, which found that 'low' vitamin D status does not reduce serum 1,25(OH)₂D concentration, and therefore calcium absorption, until the serum 25(OH)D concentration falls to around 10 nmol/L and suggested this concentration below which the formation of 1,25(OH)₂D is compromised. The Panel notes that neither Lamberg-Allardt et al. (2013), nor Newberry et al. (2014) or SACN (2016) considered the relationship between serum 25(OH)D concentration and calcium absorption.

For studies post-dating the IOM report, the Panel identified several studies, including two RCTs (Shapses et al., 2013; Aloia et al., 2014) and one observational study (Shapses et al., 2012) using the dual isotope technique to measure fractional calcium absorption. The Panel also identified two RCTs (Gallagher et al., 2012, 2014) that used a single isotope technique. They were considered as supportive evidence by the Panel and are described individually below, followed by a summary of the results and an overall conclusion on calcium absorption in adults.

With regard to results obtained with the dual isotope technique, in a 6-week placebo-controlled, double-blind RCT, Shapses et al. (2013) measured fractional calcium absorption in 83 post-menopausal women (mean age 57.8 ± 0.7 years, mean BMI of 30.2 ± 3.7 kg/m², mean baseline serum 25(OH)D concentration of 62.3 ± 14.3 nmol/L), during either a weight loss or weight maintenance period. All women were given 1.2 g calcium/day and 10 μg vitamin D₃/day, and either weekly vitamin D₃ (375 μg) or a placebo, equivalent to a total supplementation of 63 μg/day and 10 μg/day, respectively, both sufficient to maintain calcium balance. The study showed that vitamin D supplementation increases fractional calcium absorption. The Panel notes, however, that no correlation was found between fractional calcium absorption and either serum 25(OH)D or 1,25(OH)₂D concentrations at baseline or after the intervention, in this study with mean baseline serum 25(OH)D concentration of 62.3 nmol/L.

In an 8-week placebo-controlled, double-blind RCT, Aloia et al. (2014) determined fractional calcium absorption in 71 healthy women (age 58.8 ± 4.9 years; mean BMI of the groups of 26.0–27.6 kg/m², and mean baseline serum 25(OH)D concentration of 63 ± 14 nmol/L, range: 30 to > 75 nmol/L), who were assigned to placebo, 20, 50, or 100 μg/day of vitamin D₃. After adjustment for potential confounders, there was a statistically significant linear relationship between an increase in 10-week calcium absorption and increasing vitamin D₃ doses (R² = 0.41, p = 0.03) and a marginally significant linear effect by 10-week serum 25(OH)D concentration (p = 0.05, R² not reported). The changes (follow-up minus baseline) in serum 25(OH)D concentration and in calcium absorption were not significantly correlated. The Panel notes that no threshold value for serum 25(OH)D concentration in relation to calcium absorption was found in this study with final serum 25(OH)D concentrations between 40 and 130 nmol/L.

In a retrospective observational study, Shapses et al. (2012) examined the influence of body weight and hormonal and dietary factors on fractional calcium absorption in 229 adult women (age 54 ± 11 years, and BMI of 31.0 ± 7.0 kg/m²). When categorised into tertiles of BMI, mean serum 25(OH)D concentrations were significantly lower (63 nmol/L) in the obese group (mean BMI 39.0 ± 10.4 kg/m²) compared with the over- or normal weight groups (75 nmol/L) (p < 0.05), whereas mean 1,25(OH)₂D₃ concentrations were similar. Fractional calcium absorption was significantly (p < 0.05) higher in obese women compared to non-obese women. After adjustment for multiple confounders, 1,25(OH)₂D₃ was a significant predictor of fractional calcium absorption (p = 0.042), but not 25(OH)D. The Panel notes that no threshold value of 25(OH)D concentration in relation to fractional calcium absorption was found in this study.

With regard to results obtained with the single isotope technique, in a 1-year double-blind RCT, Gallagher et al. (2012) measured calcium absorption, expressed as percentage of the actual dose per litre of plasma, at baseline and 12 months in 163 post-menopausal Caucasian women (age 57–90 years) with baseline serum 25(OH)D concentrations in the range of 12.5–50 nmol/L. Participants received one of the vitamin D₃ supplementation doses of 10, 20, 40, 60, 80, 100, or 120 μg/day or placebo and mean serum 25(OH)D increased from a mean value of 38 nmol/L at baseline (all subjects) to 112 nmol/L in subjects with the highest dose of vitamin D. Calcium absorption at 12 months was more related to 12-month serum 25(OH)D concentration (R² = 0.51, p < 0.001) than to dose (R² = 0.48, p < 0.033), after adjustments for potential confounders. There was, however, no evidence for a threshold value for a reduced calcium absorption in the 12-month serum 25(OH)D concentration range of 25–165 nmol/L (values read on figure). In another 1-year double-blind RCT, Gallagher et al. (2014) measured calcium absorption (% dose per litre of plasma) at baseline and after 12 months in 198 Caucasian and African American women (age 25–45 years) with initial serum 25(OH)D concentration ≤ 50 nmol/L. Participants received a vitamin D₃ supplementation dose of 10, 20, 40, 60 μg/day or placebo and were advised to take a calcium supplement (200 mg) to maintain a calcium intake of approximately 1,000 mg/day. Mean serum 25(OH)D increased from 33.5 nmol/L (all subjects) at baseline to 100 nmol/L in the group receiving the highest dose of vitamin D₃. No changes in calcium absorption were observed over time on any dose in either Caucasians or African Americans, and no significant relationship was observed between 12-month calcium absorption and baseline or final serum 25(OH)D. No threshold value of serum 25(OH)D for calcium absorption was found at baseline or in the longitudinal study. The Panel notes that these two studies do not to identify a threshold for serum 25(OH)D concentration below which calcium absorption is impaired.

Conclusions on calcium absorption in adults

The Panel notes that all studies identified after the IOM report were conducted in women (mostly post-menopausal women), but were otherwise quite heterogeneous with respect to study design (age profile of subjects, ethnicity, body weight, dose of vitamin D, calcium supplementation), which contribute to the mixed findings and limit a conclusion. Duration of RCTs ranged between 6 weeks and 1 year.

The Panel notes that the cross-sectional single isotope study by Need et al. (2008), included in the review by the IOM, showed that calcium absorption was reduced at 25(OH)D concentrations around 10 nmol/L, below which the formation of 1,25(OH)₂D was compromised.

The Panel also notes that the two recent RCTs (Shapses et al., 2013; Aloia et al., 2014) and the one observational study (Shapses et al., 2012) using the dual isotope technique included subjects with relatively high baseline serum 25(OH)D concentrations (mean above 60 nmol/L). The Panel notes that these three studies showed no threshold value for serum 25(OH)D concentration in relation to fractional calcium absorption, in particular no threshold value in the serum 25(OH)D range between 40 and 130 nmol/L (Aloia et al., 2014) or that fractional calcium absorption was higher in the group (Shapses et al., 2012) with the lowest serum 25(OH)D concentration (mean 63 nmol/L). These results are supported by findings of two RCTs (Gallagher et al., 2012, 2014) using the single isotope technique and undertaken at lower baseline mean serum 25(OH)D concentrations (33.5 and 38 nmol/L). Results of studies are inconsistent on whether serum 25(OH)D concentration was a statistically significant predictor of calcium absorption (Gallagher et al., 2012; Aloia et al., 2014) or not.

Overall, based on these studies, the Panel considers that calcium absorption was shown to be compromised only in patients with vitamin D deficiency (with serum 25(OH)D concentration ≤ 10 nmol/L) and that the recent studies provide no evidence of a threshold effect in relation to fractional calcium absorption in adults, for serum 25(OH)D concentrations ranging between 33.5 and 75 nmol/L (mean at baseline) or between 40 to 130 nmol/L (range of final concentrations).

5.1.2.1.7 Summary of conclusions on serum 25(OH)D concentration as an indicator of musculoskeletal health in adults

The Panel notes that the evidence on a possible threshold value for serum 25(OH)D concentration with regard to the risk of adverse musculoskeletal health outcomes in adults shows a wide variability of results. Several factors contribute to this (Sections 5.1.2.1.1, 5.1.2.1.3 and 5.1.2.1.4) and also include the large variation in the results from different laboratories and assays used for measuring serum 25(OH)D concentration (Section 2.4.1). Furthermore (as indicated in the previous sections), observational studies mostly used single measurements of 25(OH)D concentration, thus possible long-term changes in 25(OH)D concentration were not considered in the analyses of the relationship with health outcomes.

The Panel concludes that, regarding the relationship between serum 25(OH)D concentration and

BMD/BMC in free-living adults, there is some evidence for a higher risk of increased BMD/BMC loss with serum 25(OH)D concentrations below 50 nmol/L,
osteomalacia, overt osteomalacia has been reported in studies on patients at mean serum 25(OH)D concentration below about 20 nmol/L, while there is some evidence that the risk of vitamin D-deficiency osteomalacia is small with serum 25(OH)D concentrations at or above 50 nmol/L,
fracture risk in free-living adults, the majority of studies indicate an increased risk of fractures associated with serum 25(OH)D concentrations of < 18 nmol/L to < 50 nmol/L,
muscle strength/function and physical performance, the evidence is inconsistent, and no target value for 25(OH)D concentration with regard to muscle strength/function and physical performance can be derived,
falls/falling, the evidence is mixed, but overall is suggestive of beneficial effects of vitamin D supplementation for reducing the risk of falls and falling in older adults over a range of serum 25(OH)D concentrations (means of 23–82 nmol/L according to the systematic reviews considered). From the available evidence, no target value for 25(OH)D concentration with regard to the risk of falls or falling can be derived,
calcium absorption, a threshold below which fractional calcium absorption is compromised has been shown in patients with serum 25(OH)D concentrations around 10 nmol/L, and that there is no evidence of a threshold effect in relation to fractional calcium absorption in adults, for serum 25(OH)D concentrations above about 30 nmol/L.

5.1.2.2 Infants and children

5.1.2.2.1 Bone mineral density/content

IOM ( 2011 ) (Section 4 and Appendix B) noted the lack of data relating serum 25(OH)D concentration to bone accretion measures in infants, and that the evidence for an association between serum 25(OH)D concentration and BMC measures in infants was inconsistent. IOM (2011) noted that, in children above 1 year of age, serum 25(OH)D concentrations of 40–50 nmol/L 'would ideally coincide with bone health benefits such as positive effects on BMC and BMD' (Viljakainen et al., 2006a; Cranney et al., 2007; Chung et al., 2009). IOM (2011) also noted that the results of RCTs in children are inconsistent when compared to results of observational studies. Overall, the IOM considered that there was some evidence for a positive association between serum 25(OH)D concentration in children and baseline BMD or change in BMD.

Lamberg-Allardt et al. (2013) based their conclusions about the possible relationship between serum 25(OH)D concentration and BMC or BMD in infants and children on Cranney et al. (2007), and their conclusions were in agreement with those derived by IOM (2011).

Newberry et al. (2014) examined the effect of vitamin D supplementation on 25(OH)D concentration and BMC in infants or children (Molgaard et al., 2010; Holmlund-Suila et al., 2012; Khadilkar et al., 2012), and considered that there was no reason to change previous conclusions (Cranney et al., 2007; Chung et al., 2009).

In infants, SACN (2016) concluded that the evidence from four intervention studies (Kim et al., 2010; Kumar et al., 2011; Abrams et al., 2012; Holmlund-Suila et al., 2012), is inconsistent with regard to an effect of vitamin D supplementation on indices of bone health in infants. SACN (2016) also noted some methodological limitations in one study (Kim et al., 2010), and the specific population of another study (undernourished low birth-weight infants (Kumar et al., 2011). For bone health indices in children aged 1–3 years, SACN (2016) identified a cross-sectional study (Hazell et al., 2015) on the relationship between plasma 25(OH)D and BMC/BMD, that is not a type of study considered by the Panel for this Section (Section 5.1.1). For children aged above 4 years, SACN (2016) concluded that a systematic review and meta-analysis including six RCTs (Winzenberg et al., 2011)26 (mean age: 10–13 years) reported a beneficial effect of vitamin D₃ supplementation on total body BMC when baseline serum 25(OH)D concentration was < 35 nmol/L. However, SACN (2016) noted that the 35 nmol/L cut-off value was arbitrarily selected based on the distribution of data (to have sufficient data for subgroup analyses). SACN (2016) also identified five trials on 'bone health indices', i.e. calcium absorption (Park et al., 2010), BMC/BMD (Ward et al., 2010; Molgaard et al., 2010; Khadilkar et al., 2012), marker of bone resorption (Ghazi et al., 2010) in children and adolescents. However, three of these studies used supplementation given monthly, bimonthly, or every third month (Ghazi et al., 2010; Ward et al., 2010; Khadilkar et al., 2012), which did not correspond to the inclusion criteria defined by the Panel for its literature search (Section 5.1.1).

The Panel retrieved five intervention and prospective observational studies, reporting on BMD/BMC in infants/children in relation to 25(OH)D concentrations and that were published after the report by IOM (2011). In the following section, the four intervention studies, first in infants then in children, are described individually, followed by the one prospective observational study in children. The results are then summarised, and an overall conclusion on BMC/BMD in infants/children is provided.

Trials with vitamin D supplementation

In a trial in 38 breastfed healthy infants (Hispanic and non-Hispanic) in the USA, who all received 10 μg/day vitamin D₃ supplementation for 3 months from 1 week after birth, Abrams et al. (2012) investigated changes in 25(OH)D concentration (cord blood then infant blood), BMC and BMD between baseline and follow-up. Mean 25(OH)D concentrations were 57.5 nmol/L (non-Hispanic) and 42 nmol/L (Hispanic) in cord blood, and were 94 nmol/L and 78 nmol/L, respectively, at age 3 months. There was no significant linear relationship between change in 25(OH)D and change in BMC. After adjustment for potential confounders, there was no significant relationship between cord 25(OH)D and BMC at 3 months. The Panel notes that, in this study of short duration (3 months), mean 25(OH)D concentration rose from about 42–58 nmol/L (cord blood) to 78–94 nmol/L at follow-up after daily vitamin D supplementation of all infants, but there was no relationship between cord 25(OH)D and BMC at 3 months.

In a double-blind randomised trial in 113 healthy term newborns (107 included in the analyses, among which 102 were breastfed infants) in Finland, Holmlund-Suila et al. (2012) investigated whether vitamin D₃ supplementation (10 μg/day or two other doses higher than the UL for infants, i.e. higher than 25 μg/day) from age 2 weeks to 3 months could ensure a serum 25(OH)D concentration of at least 80 nmol/L, without signs of excess. Samples of cord blood were collected at birth to measure baseline serum 25(OH)D, and tibia total and trabecular bone density or area, cortical bone density or area, and bone stress and strain index were assessed by pQCT (see Appendix A). Serum 25(OH)D measured at birth in cord blood did not differ among groups (mean: 52–54 nmol/L according to groups, median: 53 nmol/L in the whole population) and was 88 nmol/L (mean) at 3 months in the group receiving 10 μg/day, with a minimum value at 3 months of 46 nmol/L. After adjustment for potential confounders, there was no significant difference in bone parameters measured by pQCT between the three vitamin D-supplemented groups. The Panel notes that, in this study of short duration (2.5 months), mean serum 25(OH)D concentration rose from about 53 nmol/L (cord blood) to 88 nmol/L (in the group receiving the lowest dose) after vitamin D₃ supplementation in infants, but vitamin D₃ doses of 10 μg/day or higher did not result in differences in BMD.

In a double-blind randomised trial in Canada, 132 (mostly breast-fed) infants aged ≤ 1 month received, for 11 months, vitamin D₃ supplementation at either 10, 20, 30 or 40 μg/day (two of these doses being higher than the UL for infants, i.e. higher than 25 μg/day) (Gallo et al., 2013). The primary outcome was to achieve a plasma 25(OH)D concentration of 75 nmol/L or greater in 97.5% of infants at 3 months. Whole body and regional BMC were included among the secondary outcomes and monitored at baseline, 3, 6, 9 and 12 months of age. Mean plasma 25(OH)D concentration was 59 nmol/L (95% CI, 55–63 nmol/L) across all groups at baseline and peaked in all groups at 3 months (at 78 and 102 nmol/L in the two groups with the lowest dose). While authors reported a dose–response relationship for vitamin D dosage and plasma 25(OH)D concentration, no such relationship was observed between vitamin D dosage and BMC (lumbar spine, femur, whole body) or BMD (lumbar spine) over time. The Panel notes that, in this study, mean plasma 25(OH)D concentration rose from 59 nmol/L to at least 78 nmol/L (at 3 months) after vitamin D₃ supplementation, but vitamin D₃ doses of 10 or 20 μg/day or higher did not result in differences in BMC/BMD over 1 year.

In a double-blind RCT (Molgaard et al., 2010), 225 Danish girls (221 completers) aged 11–12 years were randomised to vitamin D₃ (5 or 10 μg/day) or placebo over 1 year with the same study design as in Viljakainen et al. (2006a) in Finnish girls (included in the review by the IOM). However, Molgaard et al. (2010) recruited the subjects during all seasons, whereas Viljakainen et al. (2006a) recruited between October and March. Whole-body and lumbar spine BMC, bone area (BA) and BMD were measured by DXA at baseline and after 12 months. Mean serum 25(OH)D (about 42–44 nmol/L) or bone measures did not differ between groups at baseline. Adjusting for baseline values, the 12-month mean change in serum 25(OH)D concentration was significantly different between groups (p < 0.0001), the final concentration being 39.7 nmol/L (−3.1 nmol/L from baseline) in the placebo group and 52.9 and 57.9 nmol/L (+11 and +13.3 nmol/L from baseline) in the 5 μg and 10 μg groups, respectively. The intervention had no effect on total body and lumbar spine BMC, BMD or BA in the whole population compared with placebo, except for the lumbar spine BA (p = 0.039, with the lowest increase in the group supplemented with 10 μg/day). The Panel notes that, in this RCT in prepubertal and pubertal girls, raising mean serum 25(OH)D concentration from 42–44 nmol/L to 53–58 nmol/L by two daily vitamin D₃ supplementation (compared with placebo) did not result in changes in BMD or BMC after 1 year.

Prospective observational study

In a UK prospective cohort study in Caucasian children (n = 2,247 in fully adjusted analyses), Sayers et al. (2012) investigated the relationship between plasma 25(OH)D₂ or 25(OH)D₃ concentrations and a number of pQCT measures (cortical BA, cortical BMC, cortical BMD, periosteal circumference, endosteal circumference and cortical thickness) (Appendix A) of the mid-tibia at age 15.5 years. Plasma 25(OH)D concentrations from samples collected at the age of 9.9 years were considered in the analysis, or at the age of 11.8 or 7.6 years if measurement at age 9.9 years was not available. Mean baseline plasma 25(OH)D₃ concentration was about 57–60 nmol/L in boys and girls, and mean 25(OH)D₂ concentration was about 4.5 nmol/L in both genders. Plasma 25(OH)D₃ concentration at baseline was significantly associated (negatively) with endosteal circumference (adjusted for periosteal circumference) and was significantly associated (positively) with cortical BMC, cortical BA or cortical thickness, after adjustment for potential confounders. The Panel notes that in this study in children with a mean baseline plasma 25(OH)D₃ concentration of about 57–60 nmol/L, plasma 25(OH)D₃ concentration was significantly associated with several bone measures.

Conclusions on BMC/BMD in infants/children

In infants, the Panel found three recent trials on BMD or BMC in (mostly) breastfed infants, two of short duration (3 months of less) (Abrams et al., 2012; Holmlund-Suila et al., 2012) and one of 11 months (Gallo et al., 2013). One trial did not show any relationship between baseline or change in mean 25(OH)D concentration (from 42–58 nmol/L (cord) up to 78–94 nmol/L) after vitamin D supplementation and BMC at 3 months (Abrams et al., 2012). After different daily doses of vitamin D supplementation, the two others did not show that increasing mean serum 25(OH)D concentrations from about 53 nmol/L (cord) (Holmlund-Suila et al., 2012) or 59 nmol/L (≤ 1 month) (Gallo et al., 2013), up to means at 3 months of at least 88 nmol/L or at least 78 nmol/L, respectively, resulted in differences in BMD/BMC (at age 3 (Holmlund-Suila et al., 2012) or 12 (Gallo et al., 2013) months).

For children, the only RCT, undertaken in prepubertal and pubertal girls, showed that raising mean serum 25(OH)D concentration from 42–44 nmol/L to 53–58 nmol/L by two daily doses of vitamin D₃ supplementation (compared with placebo) did not result in changes in BMD or BMC after 1 year (Molgaard et al., 2010). In one prospective cohort study in children with a mean baseline plasma 25(OH)D₃ concentration of about 57–60 nmol/L, plasma 25(OH)D₃ concentration was significantly associated with several bone measures (Sayers et al., 2012).

The Panel takes into account the conclusions by IOM on the relationship between serum 25(OH)D concentrations and BMC/BMD in infants (inconsistent results) and children (evidence for a positive association), and the studies published thereafter. Overall, the Panel considers that there is some evidence that, in infants and children, increasing mean serum 25(OH)D from about 40–60 nmol/L to higher values is not associated with further benefit on BMC/BMD.

5.1.2.2.2 Rickets

IOM ( 2011 ) (Section 4 and Appendix B) considered, that in the presence of an adequate calcium intake, there was evidence for an association between low mean serum 25(OH)D concentration (< 30 nmol/L) and confirmed rickets (Section 2.2.2.1) and that the risk of rickets was 'minimal when serum 25(OH)D levels range between 30 and 50 nmol/L'.

Based on Cranney et al. (2007), Lamberg-Allardt et al. (2013) concluded that there was an increased risk of rickets below a serum 25(OH)D concentration of 27.5 nmol/L, i.e. about 30 nmol/L. No new data on rickets were identified by Newberry et al. (2014). SACN (2016) concluded that the evidence from a total of 44 studies (including several case reports that is not a type of study considered by the Panel for this Section), on vitamin D and rickets is mainly observational and therefore subject to confounding. SACN (2016) notes that most studies did not report on calcium intake, thus it was unclear if rickets was caused by vitamin D deficiency or by low calcium intake or both, and that most studies did not provide information on the time of year in which the blood sample was drawn. SACN (2016) reported that serum 25(OH)D concentration in case reports ranged from < 2.5 to < 50 nmol/L and that mean/median concentrations ranged between 5 and 50 nmol/L in other study types in patients. Individual and mean serum 25(OH)D concentrations were < 25 nmol/L in the majority of studies examined.

The Panel did not find any relevant primary study on serum 25(OH)D and the risk of rickets in infants and children, providing information on their calcium intake and published after the IOM report.

The Panel takes into account the conclusions by IOM (2011) and Lamberg-Allardt et al. (2013) on evidence of overt rickets at mean serum 25(OH)D concentrations below 30 nmol/L with adequate calcium intake. Based on conclusions by IOM that the risk of rickets was minimal when serum 25(OH)D concentration ranges between 30 and 50 nmol/L, the Panel concludes that there is no risk of vitamin D-deficiency rickets with serum 25(OH)D concentrations at or above 50 nmol/L and adequate calcium intake.

5.1.2.2.3 Calcium absorption

IOM ( 2011 ) reviewed together data on calcium absorption in adults or children (Sections 4 and 5.1.2.1.6, Appendix B). The IOM concluded that, in life stages of bone accretion, maximal calcium absorption is associated with serum 25(OH)D concentrations of at least 30 nmol/L, and closer to 40–50 nmol/L, and that fractional calcium absorption does not appear to increase with serum 25(OH)D concentration above 50 nmol/L. The Panel notes that the IOM included the study by Abrams et al. (2009), which pooled studies in 251 children (about 5–17 years) using the dual isotope technique. This study found that, when serum 25(OH)D concentration was studied as a categorical variable in the whole population, fractional calcium absorption adjusted (in particular) for calcium intake was slightly but significantly higher at serum 25(OH)D concentration of 28–50 nmol/L (0.344 ± 0.019), compared with concentrations of 50–80 nmol/L (0.280 ± 0.014, p < 0.001) or greater than 80 nmol/L (0.297 ± 0.015, p < 0.007). Calcium absorption was not considered 'as such' by Lamberg-Allardt et al. (2013), Newberry et al. (2014) or SACN (2016). However, SACN (2016) considered the trial by Park et al. (2010) on fractional calcium absorption (described below).

The Panel identified one additional RCT (Abrams et al., 2013) using the dual-stable isotope technique for measuring fractional calcium absorption. As for studies on calcium absorption in adults (Section 5.1.2.1.6), the Panel also described two studies (Park et al., 2010; Lewis et al., 2013) using the single isotope technique (considered as supportive evidence by the Panel).

With regard to results obtained with the dual isotope technique, in an 8-week RCT in 63 prepubertal children aged 4–8.9 years consuming 600 to 1,200 mg/day calcium at baseline and who received 25 μg/day vitamin D₃ or a placebo (Abrams et al., 2013), mean 25(OH)D concentration was about 70 nmol/L in both groups at baseline and was significantly lower (mean ± SD: 75 ± 12 nmol/L) in the placebo than in the supplemented group (90 ± 6 nmol/L) (p = 0.01) at the end of the study period. No significant difference in fractional calcium absorption was measured at baseline and at the end of the study between the placebo group and the vitamin D₃ supplemented group. The Panel notes that, in this study, increasing mean serum 25(OH)D from 70 to 90 nmol/L by vitamin D supplementation (compared with placebo) did not result in any difference in fractional calcium absorption.

With regard to results obtained with the single isotope technique, Park et al. (2010) used a two-period metabolic balance study to investigate the effect of vitamin D supplementation on calcium absorption and retention in 11 adolescent girls aged 12–14 years with a mean baseline serum 25(OH)D concentration of 35.1 nmol/L. Subjects consumed a controlled intake (providing 5 μg vitamin D and 1,117 mg calcium/day) for two 3-week metabolic balance periods separated by a 1-week washout period. After the first metabolic balance period, participants received 25 μg/day vitamin D₃ supplementation for 4 weeks. Fractional calcium absorption was measured in each metabolic balance period using a stable calcium isotope method. All urine and faecal samples were collected and analysed to measure net calcium absorption and calcium retention. Daily supplementation with 25 μg vitamin D resulted in a mean increase in serum 25(OH)D of 13.3 nmol/L (p < 0.01) but a decrease in fractional calcium absorption of 8.3% (p < 0.05) and no significant change in fasting serum 1,25(OH)₂D, PTH, net calcium absorption, or calcium skeletal retention. The Panel notes that, in this study in pubertal girls, increasing mean serum 25(OH)D from 35.1 nmol/L to 48.4 nmol/L did not improve fractional or net calcium absorption.

In a 12-week double-blind RCT in children aged 9–13 years (165 African American and 158 Caucasian) with a mean baseline calcium intake of about 900 mg/day, Lewis et al. (2013) evaluated the effects of daily vitamin D₃ supplementation (10 μg, 25 μg, 50 μg, 100 μg) or placebo on 25(OH)D concentration and other parameters including fractional calcium absorption. Compared with a mean baseline 25(OH)D concentration of 70 nmol/L in the whole population, the mean change in 25(OH)D was −10 nmol/L for the placebo group, and ranged from +5.5 nmol/L to +76.1 nmol/L in the supplemented groups. In the whole population, 25(OH)D concentration at baseline or after 12 weeks was not related to changes in fractional calcium absorption, even after adjustment for potential confounders. There was no effect of vitamin D₃ supplementation on change in fractional calcium absorption. The Panel notes that, in this study, 25(OH)D concentration at baseline (mean: 70 nmol/L) or after 12 weeks of vitamin D supplementations compared with placebo was not related to changes in fractional calcium absorption.

Conclusions on calcium absorption in children

The Panel notes that few data are available on the relationship between serum 25(OH)D concentration and fractional calcium absorption in children.

The Panel notes that the dual isotope study by Abrams et al. (2009), included in the review by the IOM, showed that fractional calcium absorption was slightly but significantly higher at serum 25(OH)D concentration of 28–50 nmol/L (0.344 ± 0.019), compared with concentrations of 50–80 nmol/L (0.280 ± 0.014, p < 0.001) or greater than 80 nmol/L (0.297 ± 0.015, p < 0.007), among children of 5–17 years of age. The Panel also took into account a metabolic balance study in adolescent girls (Park et al., 2010) showing that increasing mean serum 25(OH)D from 35 nmol/L to 48 nmol/L did not improve fractional or net calcium absorption. In addition, the Panel notes that the two recent RCTs using the dual isotope technique (Abrams et al., 2013) or the single isotope technique (Lewis et al., 2013) in children with relatively high baseline serum 25(OH)D concentrations (mean about 70 nmol/L) did not find any relationship between fractional calcium absorption and serum 25(OH)D concentration (or any threshold value for this concentration).

Overall, based on these studies, the Panel considers that there is no relationship between fractional calcium absorption in children and serum 25(OH)D concentration above about 30–50 nmol/L.

5.1.2.2.4 Summary of conclusions on serum 25(OH)D concentration as an indicator of musculoskeletal health in infants and children

The Panel notes the paucity of data on serum 25(OH)D concentrations and musculoskeletal health outcomes in infants and children.

In spite of the large variation in the results from different laboratories and assays used for measuring serum 25(OH)D concentrations (Section 2.4.1), the Panel concludes that, regarding the relationship between serum 25(OH)D concentration and

BMD/BMC in infants and children, there is some evidence that increasing mean serum 25(OH)D from about 40–60 nmol/L to higher values is not associated with further benefit on BMC/BMD,
rickets, there is evidence of overt rickets at mean serum 25(OH)D concentrations below 30 nmol/L with adequate calcium intake, but no risk of vitamin D-deficiency rickets with serum 25(OH)D concentrations at or above 50 nmol/L and adequate calcium intake,
calcium absorption, there is no relationship between fractional calcium absorption in children and serum 25(OH)D concentration above about 30–50 nmol/L.

The Panel considers that the evidence on associations between serum 25(OH)D and musculoskeletal health outcomes is not adequate to set a different target value for serum 25(OH)D concentration in children compared to adults.

5.1.3 Serum 25(OH)D concentration and health outcomes in pregnancy

IOM ( 2011 ) (Section 4 and Appendix B) considered the following outcomes for pregnancy: calcium absorption, maternal/fetal/neonatal/childhood bone health and related outcomes (e.g. PTH), neonatal rickets, and maternal blood 25(OH)D. Separately, the IOM also considered pre-eclampsia (i.e. hypertension with proteinuria) and pregnancy-induced hypertension (i.e. transient hypertension without proteinuria). IOM (2011) concluded that calcium absorption, maternal bone health, neonatal rickets, risk of pre-eclampsia or pregnancy-induced hypertension, or non-skeletal (maternal or infant) outcomes could not be used to set DRVs for vitamin D for pregnant women. IOM concluded that fetal and childhood bone-related health outcomes were informative for the development of reference values for vitamin D in pregnancy, which in the end did not differ from that for non-pregnant women.

Newberry et al. (2014) identified one article in relation to pre-eclampsia that reported on two combined RCTs assessing the effect of supplemental vitamin D (Wagner et al., 2013b). They also refer to five nested case–control studies (Baker et al., 2010; Powe et al., 2010; Shand et al., 2010; Woodham et al., 2011; Wei et al., 2012) and two prospective cohort studies (Scholl et al., 2013; Wei et al., 2013). Newberry et al. (2014) noted that some recent studies suggest a possible relationship between vitamin D supplementation or status and the risk of preeclampsia. Newberry et al. (2014) identified two cohort studies (Bodnar et al., 2010; Burris et al., 2012) published after the report by IOM, that assessed the association between maternal serum 25(OH)D concentrations and the risk of giving birth to a small-for-gestational-age (SGA) infant (Bodnar et al. (2010) being already included in the IOM report). Newberry et al. (2014) also identified one nested case–control study and one prospective cohort study that assessed the association with preterm birth (Baker et al., 2011; Bodnar et al., 2013), of which one study was conducted in pregnant women expecting twins (Bodnar et al., 2013).

SACN (2016) identified one cohort study (Haliloglu et al., 2011) on a marker of bone turnover in pregnancy and post partum and five cohort studies (Prentice et al., 2009; Mahon et al., 2010; Viljakainen et al., 2010; Dror et al., 2012; Young et al., 2012) (some of them included in the IOM report, and some of them using predetermined cut-offs for serum 25(OH)D)). SACN (2016) reported that four of the cohort studies showed a positive association between maternal serum 25(OH)D concentration and various 'indices of bone health' in the fetus (Mahon et al., 2010; Young et al., 2012) or newborn (tibia BMC and cross-sectional area (CSA) (Viljakainen et al., 2010), or cord serum bone specific ALP and cord serum 25(OH)D (Dror et al., 2012)). SACN (2016) also reported on one multicentre double-blind randomised placebo controlled trial (Cooper et al., 2016) on maternal vitamin D₃ supplementation during pregnancy and neonatal whole body BMC (EFSA, 2016) (Section 5.2.6). SACN (2016) also considered maternal serum 25(OH)D concentration in relation to non-skeletal outcomes in the mother as well as in the newborn (risk of pre-eclampsia, neonatal hypocalcaemia, birth weight and length, risk of SGA, cognitive and psychological development in the offspring, growth and respiratory disease in the offspring). In particular, SACN (2016) considered evidence from a systematic review (Harvey et al., 2014), which reported that the association between maternal serum 25(OH)D concentration during pregnancy and pre-eclampsia and gestational diabetes is inconsistent.

The Panel undertook a literature search and also reviewed recent primary studies identified in two systematic reviews of intervention and observational studies (Harvey et al., 2014; Newberry et al., 2014). As for men and non-pregnant women and children, markers of bone formation and turnover (e.g. Haliloglu et al., 2011; Dror et al., 2012) were not an outcome considered by the Panel in view of setting DRVs for vitamin D (Section 5.1.1).

Regarding the review health outcomes in pregnancy, with the aim of setting DRVs for vitamin D:

The Panel considered available primary studies (RCTs and prospective observational studies) on serum 25(OH)D during pregnancy and maternal outcomes (bone health, for which no new data were found, pre-eclampsia or pregnancy-induced hypertension). The Panel also considered the relationship between serum 25(OH)D during pregnancy and the following outcomes in the newborn or child (but not in the fetus): bone health at birth, gestational length, anthropometry at birth in relation to the risk of SGA infant, risk of preterm birth, bone health/anthropometry/body composition after about the first year of life.
In addition, the Panel did not consider studies providing risk estimates in specific populations like women with type 1 diabetes (Azar et al., 2011), patients already with pre-eclampsia or women all recruited for being at high risk of pre-eclampsia (Shand et al., 2010; Robinson et al., 2011), or studies with supplementation of other nutrients besides vitamin D but without measurement of 25(OH)D concentration (Watson and McDonald, 2010). In addition, the Panel did not consider data on adolescent or twin pregnancies (Bodnar et al., 2013). The Panel also did not consider further investigations (Woodham et al., 2011; Wei et al., 2013) of studies described below, as these further investigations dealt with the combined association of angiogenesis and endothelial dysfunction indicators, in addition to serum 25(OH)D concentration, with the risk of preeclampsia.

The Panel identified a total of 12 references on maternal 25(OH)D concentration and: risk of pre-eclampsia, risk of being born SGA, risk of preterm birth, and bone health of the offspring.

Some studies identified considered several of these outcomes. In the following section, for each of these outcomes (Sections 5.1.3.1–5.1.3.4), the studies are described individually below; the results are then summarised, and a conclusion on maternal 25(OH)D concentration and the considered outcome is proposed. Finally, an overall conclusion for health outcomes in pregnancy is provided (Section 5.1.3.5).

5.1.3.1 Risk of pre-eclampsia

The Panel identified only two intervention studies with vitamin D during pregnancy and several outcomes including birth weight and the risk of preterm birth or pre-eclampsia, reported in one reference (Wagner et al., 2013b). The other six pertinent references on the risk of pre-eclampsia identified were observational studies and are described afterwards.

RCTs with vitamin D supplementation

Wagner et al. (2013b) combined data sets from two double-blind RCTs (Hollis et al., 2011; Wagner et al., 2013a) on healthy women (total n = 504, age ≥ 16 years) at 12 to 16 weeks of pregnancy and followed until delivery. All subjects received a prenatal 10 μg/day vitamin D₃ supplement, and were randomised to receive either a placebo, or daily doses of vitamin D₃ supplements (to reach a total intake of 50 or 100 μg/day). Serum 25(OH)D concentrations were not significantly different between groups (means between 57 and 65 nmol/L) at baseline (during pregnancy), but were higher in the supplemented groups compared to control in maternal blood within 6 weeks of delivery or in neonatal/cord blood, after adjustments for potential confounders. Four main Comorbidities Of Pregnancy (COPs), including pre-eclampsia and related hypertensive disorders as well as preterm birth without pre-eclampsia, were investigated as secondary outcomes. The study showed that the OR of any COP per 25 nmol/L increment of final maternal 25(OH)D concentration did not reach statistical significance (but the risk was significantly reduced when all COPs were considered together). Neonatal birth weight did not significantly differ between supplemented groups and controls. The Panel notes that there was no effect of daily supplementation with vitamin D₃ during pregnancy on neonatal birth weight, and risk of pre-eclampsia or preterm birth in this population with mean serum 25(OH)D concentrations of 57–65 nmol/L at baseline.

Prospective observational studies

In the following observational studies, pre-eclampsia was defined as the occurrence of gestational hypertension in previously normotensive women accompanied by new-onset proteinuria after 20 weeks of gestation. Definition of pre-eclampsia based on values of systolic and/or diastolic blood pressure and proteinuria, although close, differed between studies, and severe pre-eclampsia was defined based on higher values of systolic blood pressure/diastolic blood pressure or proteinuria.

In a nested case–control study in the USA, Powe et al. (2010) assessed the association between first trimester total serum 25(OH)D concentrations and development of pre-eclampsia in 39 cases (with a significantly higher first trimester systolic and diastolic blood pressure), and 131 normotensive control women (who remained normotensive in pregnancy, did not have gestational diabetes mellitus or did not give birth to SGA infants). Baseline serum 25(OH)D concentrations did not differ significantly between cases and controls (mean about 68 and 72 nmol/L, respectively, measured at (mean ± SD) 11.2 ± 3.6 versus 11.6 ± 3.0 weeks of gestation) and were not associated with baseline systolic or diastolic blood pressure. No association was found between first trimester serum 25(OH)D concentration (per 25 nmol/L increase, across quartiles, or for those < or > 37.5 nmol/L) and risk of subsequent pre-eclampsia, after full adjustments for potential confounders. The Panel notes that this study did not report an association between serum 25(OH)D concentration during the first trimester of pregnancy and incidence of pre-eclampsia.

One nested case–control study by Baker et al. (2010) was conducted in the USA in a population selected from a cohort of 3,992 healthy women, who had previously given blood in the framework of routine prenatal care. The study analysed maternal 25(OH)D status during mid-gestation (15–20 weeks of gestation) and risk of development of severe pre-eclampsia. From the cohort, a case group of 51 women was identified who developed severe pre-eclampsia (median age 28 years), out of which 41 women were included in the analysis. The control group was composed of 198 randomly selected ethnicity-matched healthy women delivering at term. Median serum 25(OH)D concentration in the case group was 75 nmol/L, which was significantly lower than that in the control group, i.e. 98 nmol/L. After adjustment for potential confounders, the risk of severe pre-eclampsia in women with mid-gestation 25(OH)D concentration of less than 50 nmol/L (n = 19 controls and 11 women with severe pre-eclampsia) was fivefold higher (OR: 5.41; 95% CI: 2.02–14.52) than in women with mid-gestation 25(OH)D of at least 75 nmol/L (n = 138 controls and 22 women with severe pre-eclampsia). There was no significant difference in risk in women with 25(OH)D between 50 and 74.9 nmol/L (n = 41 controls, and 10 with severe pre-eclampsia) compared with 25(OH)D of at least 75 nmol/L. The Panel notes that this study found that the risk for severe pre-eclampsia was higher in women with a 25(OH)D concentration at 15–20 weeks of gestation less than 50 nmol/L in comparison to those with concentrations higher than 75 nmol/L.

In a case–control study in the USA, Robinson et al. (2010) investigated maternal plasma 25(OH)D concentration in 50 women with diagnosed early onset severe pre-eclampsia (EOSP, diagnosed before 34 weeks of gestation) compared to 100 ethnicity- and gestational age-matched healthy controls followed throughout their normal singleton pregnancy. Plasma 25(OH)D concentration (median, interquartile range (IQR)), was obtained in the cases at time of diagnosis (45, 32.5–77.5 nmol/L) and was significantly lower than in controls (80, 50–110 nmol/L; p < 0.001), both at a mean gestational age of 29 weeks (28–31 weeks in cases, 26–31 weeks in controls). Birth weight and gestational age at delivery were significantly lower in cases than in controls, while mean arterial pressure at sample collection and incidence of intrauterine growth restriction (i.e. less than 10th percentile birth weight for gestational age) were significantly higher. After adjustment for potential confounders, there was a significant association between a 25 nmol/L increase in maternal plasma 25(OH)D and a reduced risk of EOSP (OR: 0.37; 95% CI: 0.22–0.62, p < 0.001). Women with plasma 25(OH)D concentration ≤ 49 nmol/L (lowest quartile) had a 3.6-fold increased risk of EOSP compared to women with higher concentrations (OR: 3.60; 95% CI: 1.71–7.58, p < 0.001). The Panel notes that this study indicates that the risk for EOSP was 3.6 times higher in women with a plasma 25(OH)D concentration less than 50 nmol/L at about 34 weeks of gestation in comparison with women with higher plasma 25(OH)D concentrations.

In a Spanish prospective cohort study in unsupplemented women followed from pregnancy to delivery (n = 466 at delivery), Fernandez-Alonso et al. (2012) investigated the relationship between first-trimester serum 25(OH)D concentration and obstetric and neonatal pregnancy outcomes. These included pre-eclampsia, gestational hypertension, preterm birth (i.e. birth at 21–37 weeks of pregnancy), and number of SGA infants (i.e. with birth weights below the 10th percentile for gestational age). Serum 25(OH)D concentration at 11–14 weeks of pregnancy was below 50 nmol/L, between 50 and 74 nmol/L or at least 75 nmol/L for, respectively, 109, 191 and 166 women. No significant non-parametric correlations were found between the first-trimester 25(OH)D levels and several numeric obstetric or neonatal outcome variables. The Panel notes that this study only assessed correlations between 25(OH)D concentrations and obstetric or neonatal outcomes.

In post-hoc analyses on a group of 697 Canadian women, who had previously participated during pregnancy in a multicentre trial of vitamin C and E supplementation and prevention of pre-eclampsia, Wei et al. (2012) measured plasma 25(OH)D concentration in maternal blood samples collected during that previous trial at visit 1 (baseline, 12–18 weeks of gestation) and visit 2 (24–26 weeks of gestation). The purpose of these new analyses was to investigate the association between maternal 25(OH)D concentrations and risk of pre-eclampsia. In the previous trial, the subjects had been stratified in the 'high-risk' group (n = 229, with at least one of four risk factors for pre-eclampsia identified by the authors) or in the 'low-risk' group (n = 468, women who had been nulliparous without risk factors for pre-eclampsia). The difference between maternal mean 25(OH)D concentrations in pre-eclamptic (n = 32) and non-pre-eclamptic (n = 665) women was not statistically significant at visit 1 (about 51–56.0 nmol/L), but significant at visit 2 (mean ± SD: 48.9 ± 16.8 nmol/L versus 57.0 ± 19.1 nmol/L, p = 0.03), in particular in the low-risk group. After adjustments for potential confounders (including the risk group), the risk of pre-eclampsia associated with maternal 25(OH)D < 50 nmol/L at 24–26 weeks of gestation (n = 236, including 19 pre-eclamptic) was 3.2-fold higher (OR: 3.24; 95% CI: 1.37–7.69) compared with maternal 25(OH)D ≥ 50 nmol/L (n = 368, 9 pre-eclamptic). This relationship was not observed for maternal 25(OH)D < 50 nmol/L (n = 272, 15 pre-eclamptic) or ≥ 50 nmol/L (n = 425, 17 pre-eclamptic) earlier in pregnancy, i.e. at 12–18 weeks of gestation. The Panel notes that according to these study findings, the risk of pre-eclampsia associated with maternal 25(OH)D concentration < 50 nmol/L at 24–26 weeks of gestation (but not at 12–18 weeks) was significantly higher compared with maternal 25(OH)D ≥ 50 nmol/L.

In a prospective cohort study on 1,141 US healthy pregnant women (mainly Hispanic and African American), Scholl et al. (2013) analysed the association of serum 25(OH)D concentration < 50 nmol/L (with or without PTH > 6.82 pmol/L) and the risk of pre-eclampsia. Maternal serum 25(OH)D concentration was measured at (mean ± SD) 13.7 ± 5.7 weeks of gestation, as 25(OH)D₃ and 25(OH)D₂, but mean baseline value was not reported. About 6% of women developed pre-eclampsia. After adjustment for potential confounders, and compared with women with 25(OH)D concentration of at least 50 nmol/L (n = 750), the risk of pre-eclampsia was significantly twofold higher in pregnant women with concentrations lower than 30 nmol/L or between 30 and 39 nmol/L (n = 121 and 116, respectively, e.g. adjusted OR for 25(OH)D < 30 nmol/L: 2.13; 95% CI: 1.07–4.26, p for trend = 0.027) (but the risk was not significantly reduced in the 154 women with 25(OH)D of 40–50 nmol/L). Women with secondary hyperparathyroidism (n = 72, PTH > 6.82 pmol/L and serum 25(OH)D < 50 nmol/L) had a 2.8-fold increase in risk (95% CI: 1.28–6.41) compared to women with similar 25(OH)D concentration without 'high' PTH. The Panel notes that, according to this cohort study in mainly Hispanic and African American women, the risk of pre-eclampsia was about twofold higher when the 25(OH)D concentration of the mother at 13.7 ± 5.7 weeks of gestation was < 40 nmol/L compared to those with a concentration ≥ 50 nmol/L.

Conclusions on risk of pre-eclampsia

The Panel notes that an increase in serum 25(OH)D concentration from a mean baseline of 57–65 nmol/L (after vitamin D supplementation in the second trimester of pregnancy compared with placebo) did not result in a change in the risk of pre-eclampsia (Wagner et al., 2013b). Out of six observational studies, two (Powe et al., 2010; Fernandez-Alonso et al., 2012) found no association between serum 25(OH)D during pregnancy (at time points of about 11–14 weeks of gestation), and risk of pre-eclampsia. In these two studies, investigated (predefined) cut-offs for 25(OH)D were < 37.5 and 50 nmol/L (versus > 37.5 or > 75 nmol/L). In contrast, four observational studies (Baker et al., 2010; Robinson et al., 2010; Wei et al., 2012; Scholl et al., 2013) found a significant association between low maternal serum 25(OH)D concentration (measured between about 13 to 31 weeks of gestation) and risk of pre-eclampsia or severe pre-eclampsia. In these studies, the investigated cut-offs, often predefined, were < 30 nmol/L, 30–39 nmol/L or < 50 nmol/L, compared most often with > 50 nmol/L (or ≥ 75 nmol/L). Overall, the Panel considers that the evidence of an association between maternal serum 25(OH)D concentration and risk of pre-eclampsia is inconsistent, although there is some evidence suggestive of an increase in the risk of pre-eclampsia at 25(OH)D concentrations below about 50 nmol/L.

5.1.3.2 Risk of being born small-for-gestational-age

With regard to the risk of being born SGA, the Panel considered four observational studies, including the study by Fernandez-Alonso et al. (2012) mentioned above.

Prospective observational studies

In a prospective population-based cohort study on healthy Danish Caucasian women (Moller et al., 2012), the association between preconception 25(OH)D concentration and several outcomes was investigated. Outcomes included incidence of miscarriage and birth outcomes, i.e. birth weight and length, head circumference, number of SGA infants, in 153 women with immediate pregnancy plans (whose vitamin D status was compared to 75 women (50 completers) who had no pregnancy plans for the next 21 months as age-matched controls). Plasma 25(OH)D concentration was measured in both groups on four occasions (at baseline, and once at each of the follow-up visits every trimester). Median (IQR) baseline plasma 25(OH)D concentration (70 (56–92) nmol/L) was significantly (p < 0.001) higher in the control group compared to women with pregnancy plans (59 (46–71) nmol/L). Baseline mean plasma 25(OH)D concentrations did not differ between those who experienced miscarriage (n = 8, out of 92 who conceived) and those who did not. Plasma 25(OH)D concentration (at baseline, at each visit, or on average during pregnancy) was not associated with gestational length, birth weight, birth length, head circumference, incidence of SGA infants, even after adjustments for potential confounders. The Panel notes that this study, in a population with baseline median plasma 25(OH)D concentration of about 50–70 nmol/L, did not find an association between maternal 25(OH)D concentration during pregnancy and anthropometric outcomes in the newborn or SGA incidence.

In a prospective cohort study of pregnant women in the US, Burris et al. (2012) assessed the association between second trimester maternal plasma 25(OH)D concentration (947 Caucasians, 186 African Americans) or cord plasma 25(OH)D concentration (606 Caucasians, 128 African Americans) and the risk of being born SGA. Women were included at less than 22 weeks of singleton pregnancies. Mean ± SD maternal and cord 25(OH)D concentrations were 60 ± 21 (at 26–28 weeks of gestation) and 47 ± 19 nmol/L, respectively, and there were 53 SGA infants. After adjustments for potential confounders, maternal or cord plasma 25(OH)D < 25 nmol/L was associated with a significantly increased risk of being born SGA, compared with plasma 25(OH)D of 25 nmol/L or greater. Indeed, the adjusted OR of being born SGA was 3.17 (95% CI: 1.16–8.63) for maternal plasma < 25 nmol/L (7 SGA infants from mothers in this category), and 4.64 (95% CI: 1.61–13.36) for cord plasma < 25 nmol/L (9 SGA infants in this category). The Panel notes that this study in second trimester pregnant women showed that maternal or cord plasma/serum 25(OH)D concentrations below 25 nmol/L (versus at least 25 nmol/L) were associated with increased risk of being born SGA.

In a US prospective cohort study, Gernand et al. (2013) studied 2,146 pairs of singleton term newborns and mothers (52% Caucasian, with no pre-existing diabetes or hypertension) who had participated in a large multicentre observational study (63% study sites at latitude ≥ 41° North). The aim of the study was to investigate the association between maternal 25(OH)D concentration and several outcomes, including the risk of being born SGA. Maternal serum 25(OH)D concentration was measured at 26 weeks of gestation or less, and every 8 weeks afterwards (mean baseline: 51.3 ± 28.0 nmol/L). There were 395 SGA infants. After adjustments for potential confounders, the risk of being born SGA was half in infants whose mothers had first trimester 25(OH)D of ≥ 37.5 nmol/L, compared to < 37.5 nmol/L (OR: 0.50; 95% CI: 0.27–0.91) (11.8 and 23.8% of SGA infants from mothers in each category). This association was not observed in the second trimester. The Panel notes that this study showed that maternal serum 25(OH)D concentrations above 37.5 nmol/L in the first trimester of pregnancy, but not the second trimester, were associated with half the risk of SGA compared with serum concentrations below 37.5 nmol/L.

Conclusions on risk of being born SGA

The Panel notes that, in contrast to Fernandez-Alonso et al. (2012) and Moller et al. (2012) (which measured frequency), two larger observational studies (Burris et al., 2012; Gernand et al., 2013) using predefined 25(OH)D cut-off values found an association of maternal 25(OH)D < 25 nmol/L (at 26–28 weeks of gestation) or < 37.5 nmol/L (in the first trimester, but not the second) with an increased risk of being born SGA (versus higher values). The Panel concludes that the evidence of an association between maternal serum 25(OH)D concentration and risk of being born SGA is inconsistent, although there is some evidence suggestive of an increase in the risk at 25(OH)D concentrations below about 25–37.5 nmol/L.

5.1.3.3 Risk of preterm birth

With regard to the risk of preterm birth, in addition to the two intervention studies reported in one reference (Wagner et al., 2013b) already described above (Section 5.1.3.1), the Panel identified one nested case–control study.

Baker et al. (2011) assessed the relationship between maternal 25(OH)D concentration during pregnancy and the risk of preterm birth in a US nested case–control study of 4,225 women with singleton pregnancies, from whom blood had been collected at 11–14 weeks of gestation for the screening of trisomy 21. Preterm birth was defined as spontaneous delivery between 23 and 35 weeks of gestation. Fourty women with preterm birth were compared to ethnicity-matched randomly selected healthy controls who delivered at term (n = 120) and gave blood at a similar gestational age. Median (IQR) serum 25(OH)D concentration for the whole study group was 89 (73–106) nmol/L. After adjustment for potential confounders, there was no association between maternal serum 25(OH)D concentration (< 50 nmol/L or 50–74.9 nmol/L, compared with ≥ 75 nmol/L) and the risk of preterm birth. The Panel notes that this study found no association between 25(OH)D concentration during pregnancy and the risk for preterm birth in this population with high baseline median 25(OH)D value (about 90 nmol/L).

5.1.3.4 Bone health of the offspring

With regard to bone health of the offspring, the Panel considered one observational study.

Viljakainen et al. (2011) evaluated in a Finnish prospective cohort study, whether there was a catch-up in tibia BMC or CSA in children (n = 87) at 14 months, from a group of 125 children previously assessed at birth (Viljakainen et al., 2010). These infants had been categorised according to maternal vitamin D status during pregnancy (defined as the mean of the first-trimester and of the 2-day post-partum serum 25(OH)D concentrations below or above the median of 42.6 nmol/L). BMD, BMC and CSA of the left tibia were measured in the newborns and at 14 months by pQCT (Appendix A). Complete baseline and follow-up data were available for 29 and 26 children whose mothers had, respectively, lower or higher vitamin D status during pregnancy. Whereas tibia BMC at birth was significantly higher in children whose mothers had a high (i.e. above median) vitamin D status during pregnancy (Viljakainen et al., 2010), the mean total BMC gain over 14 months was significantly higher in the children whose mothers had a low vitamin D status (0.062 g/cm², p = 0.032) resulting in similar BMC in both groups of children at 14 months (Viljakainen et al., 2011). Although tibia CSA at birth was significantly larger in children whose mothers had a high vitamin D status during pregnancy (Viljakainen et al., 2010), the differences between groups in mean CSA change over 14 months or in final CSA at 14 months did not reach statistical significance (Viljakainen et al., 2011). The Panel notes that maternal 25(OH)D at or below about 43 nmol/L during pregnancy was associated with bone outcomes in the child at birth, which did not persist at the age of about 1 year possibly due to infant vitamin D supplementation starting at 2 weeks of age.

5.1.3.5 Summary of conclusions on serum 25(OH)D concentration and health outcomes in pregnancy

The Panel notes that the evidence on a possible threshold value for serum 25(OH)D concentration with regard to the risk of adverse pregnancy-related health outcomes shows a variability of results. Several factors contribute to this (as also discussed in Sections 5.1.2.1.1, 5.1.2.1.3 and 5.1.2.1.4 for musculoskeletal health outcomes in adults) and also include the large variation in the results from different laboratories and assays used for measuring serum 25(OH)D concentrations (Section 2.4.1). Furthermore, observational studies often used single measurements of 25(OH)D concentration, thus possible changes in 25(OH)D concentration throughout pregnancy were not considered in the analyses of the relationship with health outcomes.

The Panel concludes that, regarding the relationship between maternal serum 25(OH)D concentration and

pre-eclampsia, there is inconsistent evidence of an association between maternal serum 25(OH)D concentration and risk of pre-eclampsia or severe pre-eclampsia, but that there is some evidence suggesting an increase in the risk at 25(OH)D concentrations below about 50 nmol/L.
risk of SGA, there is inconsistent evidence of an association of maternal 25(OH)D concentration with an increased risk of SGA, but that there is some evidence suggesting an increase in the risk at 25(OH)D concentrations below about 25–37.5 nmol/L.
risk of preterm birth, there is no evidence of an association.
indicators of bone health in the child after birth, although maternal 25(OH)D at or below about 43 nmol/L during pregnancy was associated with bone outcomes in the child at birth, there is no evidence of an association persisting at the age of about 1 year.

5.1.4 Serum 25(OH)D concentration and health outcomes in lactation

IOM ( 2011 ) (Section 4 and Appendix B) noted that maternal serum 25(OH)D concentrations increased after vitamin D supplementation of lactating mothers, but that this supplementation had no significant effect on either infant serum 25(OH)D concentrations (for supplementation below 100 μg/day) or infant weight or height. The IOM also noted that there was a lack of association between maternal 25(OH)D concentration and maternal post partum changes in BMD, or breast milk calcium content. The IOM considered that neither maternal BMD nor maternal or fetal serum 25(OH)D concentrations could be used to set reference values for vitamin D during lactation.

SACN (2016) considered one review on vitamin D supplementation during lactation in relation to breast milk vitamin D concentration and serum 25(OH)D concentration in exclusively breast-fed infants (Thiele et al., 2013) and stated that the vitamin D concentration of breast milk increased significantly following supplemental vitamin D of ≥ 50 μg/day but not of 10 μg/day. SACN (2016) also discussed data on plasma 25(OH)D concentration and vitamin D and 25(OH)D concentration in breast milk from lactating mothers (við Streym et al., 2016).

The Panel undertook a literature search to identify primary studies (RCTs and prospective or case–control observational studies) on the relationship between maternal serum 25(OH)D and health outcomes of mothers during lactation, published after the evidence reviewed by IOM (2011). The Panel also considered the systematic review by Newberry et al. (2014). In its search, as for pregnancy-related outcomes (Section 5.1.3), the Panel did not consider data on lactating adolescent. The Panel identified one study published in 2010 on the relationship between maternal serum 25(OH)D and health outcomes of lactating women that is described hereafter.

Salama and El-Sakka (2010) assessed vitamin D in a cohort of 32 breastfed infants (exclusively (n = 20) or partially) with rickets (including nine with hypocalcaemic seizures) and their lactating mothers, in Egypt. Subjects were identified based on clinical presentation, biochemical results and radiological findings, and serum concentrations of calcium, phosphorus, ALP, 25(OH)D and PTH were measured (calcium intake was not reported). Neither infants or their mothers received calcium or vitamin D supplementation and all had limited sun exposure. Infants were aged (mean ± SD) 3.7 ± 1.6 months or 12.4 ± 4.3 months, in the groups with or without hypocalcaemic seizures, respectively. Median (IQR) serum 25(OH)D concentration was 40 (45) nmol/L in mothers (range 10–175 nmol/L), and was 37.5 (32.5) nmol/L in infants (range 7.5–95 nmol/L), with median (IQR) of 17 (25) and 45 (25) nmol/L in the groups with or without hypocalcaemic seizures, respectively. The correlation between serum 25(OH)D concentrations in rachitic infants and serum 25(OH)D concentrations in their mothers (r = 0.326) was not statistically significant. The Panel notes that this study found no significant association between serum 25(OH)D concentrations in infants with rickets and in their lactating mothers.

Conclusions on serum 25(OH)D concentration and health outcomes in lactation

The Panel notes that the only recent study identified by the Panel found no significant association between serum 25(OH)D concentrations in infants with rickets and serum 25(OH)D concentrations in their mothers. Data on the low concentration of vitamin D in breast milk, and on vitamin D intake and status of lactating women were discussed by the Panel previously (Section 2.3.7.2).

The Panel concludes that there is no evidence for a relationship between serum 25(OH)D concentration and health outcomes of lactating women that may be used to set a DRV for vitamin D.

5.1.5 Serum 25(OH)D concentration and non-musculoskeletal health outcomes

For non-musculoskeletal health outcomes, as indicated in Section 5.1.1, the Panel considered the evidence collated in and conclusions of the report by IOM (2011), the systematic review by Newberry et al. (2014) and the report by SACN (2016). The Panel's main objective in this section was to investigate whether data on serum 25(OH)D concentration and non-musculoskeletal health outcomes may be used to set a target value for serum 25(OH)D in order to derive DRVs for vitamin D. As the three reports the Panel considered may have had different objectives (e.g. without always drawing separate conclusions for vitamin D intake and vitamin D status), the overall conclusions of these reports with regard to the relationship between vitamin D intake (either alone or with calcium) or status (i.e. serum 25(OH)D concentration) and several health outcomes are briefly summarised below.

The three reports covered often the same health outcomes (cancer, cardiovascular diseases (CVD), markers of immune function, function of the nervous system and risk of related disorders, non-skeletal obstetric outcomes), with some exceptions. For example, all-cause mortality and pancreatic cancer were covered by Newberry et al. (2014) and not by IOM (2011). Type 2 diabetes and metabolic syndrome, functions of the nervous system and risk of related disorders (e.g. cognition, mood, depression, autism) and non-skeletal obstetric outcomes were covered by IOM (2011) (Appendix B) and not by Newberry et al. (2014). Other cancers (such as oesophagus, stomach, larynx, oropharynx, lung, endometrium, ovary, kidney, non-Hodgkin, liver, bladder, melanoma and basal cell skin cancer), maternal serum 25(OH)D concentration in pregnancy and later cognitive and psychological development of the offspring, neonatal hypocalcaemia, oral health and age-related macular degeneration (AMD) were only covered by SACN (2016).

According to these reports, there is no or an inconsistent relationship between vitamin D intake (with or without calcium) or status and all-cause mortality or total cancer risk and mortality, although SACN (2016) reported conclusion from a systematic review that vitamin D supplementation in combination with calcium reduces mortality risk and that this is not seen with vitamin D supplementation alone. Most of the evidence on breast cancer, colorectal cancer and prostate cancer, was of observational nature and was considered of limited value or inconsistent or insufficient to conclude on a dose–response relationship. However, Newberry et al. (2014) concluded that the only observational evidence identified in their update for pancreatic cancer found an increase in the risk with increased serum 25(OH)D concentration.

For total CVD/cardiovascular events and hypertension, IOM (2011), Newberry et al. (2014) and SACN (2016) concluded that no or an inconsistent relationship was found between vitamin D intake (with or without calcium) or status and the risk of these outcomes, based on evidence which was considered limited, not statistically significant or not supported by intervention studies. However, when addressing CVD mortality separately, Newberry et al. (2014) concluded that 8 observational studies (prospective cohort or nested case–control studies, no RCTs) showed a higher risk for cardiovascular death for subjects with the lowest serum 25(OH)D concentrations (lower bounds throughout all the studies ranged between 8 and 40 nmol/L) compared with those with the highest (higher bounds ranged between 45 and > 100 nmol/L).

The evidence on type 2 diabetes and metabolic syndrome was considered not conclusive by the IOM for the purpose of setting DRVs. In addition, limited or inconsistent evidence of mostly observational nature was also found on the relationship between vitamin D intake (either alone or with calcium) or status and functions of the nervous system and the risk of related disorders.

For markers of immune function, IOM (2011), Newberry et al. (2014) and SACN (2016) considered a variety of outcomes including asthma, autoimmune diseases, wheeze, atopy and various infectious diseases and IOM (2011) and SACN (2016) concluded that the evidence for a cause and effect relationship was insufficient for setting DRVs for vitamin D.

For non-skeletal obstetric outcomes (caesarean section, obstructed labour in the mother, and immune-related outcomes in the offspring such as type 1 diabetes mellitus, asthma and atopic disorders, or other outcomes in the offspring, e.g. Apgar score assessed by one and/or the other report), IOM (2011) and SACN (2016) concluded that the evidence is limited and not conclusive, as conflicting results are shown in observational studies and RCTs.

For all the health outcomes (other cancers, maternal serum 25(OH)D concentration in pregnancy and later cognitive and psychological development of the offspring, neonatal hypocalcaemia, oral health, AMD) assessed only by SACN (2016), the evidence from observational studies is not supported by robust clinical trials or evidence is lacking, or inconsistent, or only weak.

The Panel also noted the large systematic review of prospective cohort studies and randomised trials by Autier et al. (2014) whose conclusions are in line with those from IOM (2011), Newberry et al. (2014) and SACN (2016), in that many prospective studies have shown associations between low 25(OH)D concentrations and a wide range of health disorders, but that a similar number of RCTs did not provide evidence for a causal relationship between serum 25(OH)D concentrations and the occurrence or the course of such disorders.

The Panel notes the consistency of the overall conclusions of the systematic reviews and reports discussed in this Section with regard to non-musculoskeletal health outcomes, thus did not undertake a specific literature search for these outcomes in view of setting DRVs for vitamin D. The Panel considers that the available evidence on these non-musculoskeletal health outcomes is insufficient to be used as criterion for setting DRVs for vitamin D.

5.1.6 Overall conclusions on serum 25(OH)D concentration and various health outcomes, in relation to the setting of DRVs for vitamin D

The Panel notes that most evidence on the relationship between serum 25(OH)D concentration and health outcomes is related to musculoskeletal health outcomes (Section 5.1.1). The Panel notes that the evidence on a possible threshold value for serum 25(OH)D concentration with regard to adverse musculoskeletal or pregnancy-related health outcomes, that may be used to inform the setting of DRVs for vitamin D, shows a wide variability of results (Sections 5.1.2.1.7, 5.1.2.2.4 and 5.1.3). Several factors contribute to this (Sections 5.1.2.1.1, 5.1.2.1.3 and 5.1.2.1.4) and also include the large variation in the results from different laboratories and assays used for measuring serum 25(OH)D concentrations (Section 2.4.1). Furthermore, observational studies mostly used single measurements of 25(OH)D concentration, thus possible long-term changes in 25(OH)D concentration were not considered in the analyses of the relationship with health outcomes.

Taking into account the overall evidence and uncertainties for adults (Section 5.1.2.1.7) and infants and children (Section 5.1.2.2.4), the Panel considers that there is sufficient evidence for an increased risk of adverse musculoskeletal health outcomes at 25(OH)D concentration below 50 nmol/L. Taking into account the overall evidence and uncertainties for pregnancy (Section 5.1.3), the Panel considers that there is also evidence for an increased risk of adverse pregnancy-related health outcomes at 25(OH)D concentration below 50 nmol/L. The Panel concludes that this concentration can be used as a target value to derive a DRV for vitamin D intake for adults, infants, children and pregnant women. The setting and analyses of the available studies do not allow a conclusion to be drawn as to whether this concentration should be achieved by about half of or most subjects in the population.

The Panel notes that there is no evidence for a relationship between serum 25(OH)D concentration and health outcomes of lactating women that may be used to set a DRV for vitamin D.

5.2 Vitamin D intake from supplements and musculoskeletal health outcomes, pregnancy and lactation

Following a similar approach as in Section 5.1 for serum 25(OH)D concentration and health outcomes, the Panel considered studies (here, preferably RCTs) on vitamin D intake (mostly as supplements, with or without calcium) and various health outcomes (several musculoskeletal health outcomes in children and adults (including free-living older adults), health outcomes in pregnancy and lactation, as defined in Section 5.1), to evaluate whether they might inform the setting of DRVs for vitamin D.

5.2.1 Bone mineral density/content in adults

IOM ( 2011 ) (Sections 4 and 5.1.2.1.1, Appendix B) reported that most of the studies (all expect one of the 18 RCTs cited) evaluated the effect of vitamin D supplementation in combination with calcium supplementation, often without information on the habitual dietary intake from foods (eight RCTs). These RCTs were predominantly conducted in post-menopausal women, using supplemental vitamin D at doses of 7.5–25 μg/day (all except two RCTs), along with 377–1,450 mg/day of calcium. From these RCTs, the IOM concluded that there was evidence that supplementation of vitamin D plus calcium (compared with placebo) resulted in small increases in BMD of the spine, total body, femoral neck and total hip, but that the evidence on vitamin D supplementation alone and BMD was limited. SACN (2016) (Section 5.1.2.1.1) concluded that the evidence was suggestive of a beneficial effect of vitamin D supplementation on bone health indices at some skeletal sites in adults aged ≥ 50 years, but that the evidence for adults < 50 years was insufficient to draw conclusions.

The Panel takes into account the same seven RCTs that were considered in relation to associations between serum 25(OH)D concentrations and BMD/BMC, from which only two (Macdonald et al., 2013; Uusi-Rasi et al., 2015) provided data on vitamin D intake from food (and possible supplements other than that of the intervention) in the study population (Section 5.1.2.1.1). The Panel notes that two of the seven RCTs found no effect on BMD of vitamin D plus calcium, from supplements or fortified foods, at doses of about 71 μg/day (Jorde et al., 2010) or 20 μg/day (Kukuljan et al., 2011), in subjects with mean baseline 25(OH)D concentrations of 58 and 86 nmol/L, respectively.

In contrast, three RCTs (Section 5.1.2.1.1) in subjects with mean baseline concentrations of 25(OH)D of 34–50 nmol/L reported an increase in BMD or a decrease in BMD loss following vitamin D supplementation at doses of 10–25 μg/day (with or without calcium) (Islam et al., 2010; Kärkkäinen et al., 2010; Macdonald et al., 2013) (results from unadjusted analyses in (Kärkkäinen et al., 2010)). One RCT (Nieves et al., 2012) in subjects with mean baseline concentration of 29 nmol/L found an increase in BMD following vitamin D supplementation with 25 μg/day plus calcium only in subjects with the FF genotype (but not in subjects with the Ff/ff Fok1 genotypes). One RCT (Uusi-Rasi et al., 2015) in subjects with mean baseline concentration of 66 nmol/L found that vitamin D supplementation of 20 μg/day alone (above the usual intake of 10 μg/day) resulted in a lower decrease in femoral neck BMD as compared to placebo, but had no effect on BMD at other sites. The controls (to which the intervention was compared to) in these studies were of various nature (Section 5.1.2.1.1).

For the present Section, the Panel also identified one prospective observational study in 9,382 women and men in Canada aged 25 years to more than 71 years and followed for 10 years, that investigated changes over time in calcium and vitamin D intakes (from foods and supplements, assessed repeatedly by food frequency questionnaires (FFQs)), and their longitudinal associations with BMD (Zhou et al., 2013). The Panel notes that, in this study, after adjustments for potential confounders, vitamin D intakes ≥ 10 μg/day (mean 10-year intake) were positively associated with 10-year BMD change at total hip or femoral neck, compared with intakes of vitamin D < 5 μg/day, in women (but not in men) (e.g. for total hip: 0.008 g/cm²; 95% CI: 0.003–0.013).

The Panel notes that the results of these studies with heterogeneous designs are not consistent. In line with the conclusions of the report by IOM (2011), altogether, the Panel notes that there is some evidence suggesting that beneficial effects of vitamin D supplementation on BMD/BMC may be achieved with doses of about 10 to 25 μg/day in free-living subjects with mean 25(OH)D concentrations between 29 and 50 nmol/L, and that the effects may depend on calcium intake.

5.2.2 Fracture risk in adults

IOM ( 2011 ) (Sections 4 and 5.1.2.1.3, Appendix B) reviewed a total of 19 RCTs identified by Cranney et al. (2007) (15 RCTs), Chung et al. (2009) (two RCTs) or by additional literature searches (2 RCTs). These RCTs provided vitamin D₂ or D₃ (with or without calcium), with various doses (e.g. out of the 15 RCTs identified by Cranney et al. (2007), 11 used vitamin D₃ doses of 7.5–20 μg/day), at various frequency (e.g. daily, every 4 months, once per year), and often with no information on the habitual dietary intake of vitamin D from foods. The IOM concluded that vitamin D supplementation with calcium was effective in reducing fracture risk (total or hip) in institutionalised older populations only (considering a limited number of studies out of the 15 RCTs identified by Cranney et al. (2007)), but that the evidence for a benefit of vitamin D and calcium supplementation on fracture risk in community-dwelling individuals was inconsistent across trials.

Newberry et al. (2014) identified one RCT using vitamin D and calcium, that assessed fracture risk, and that was not already considered by the IOM. This RCT (Prentice et al., 2013) was a re-analysis of data from a previous trial that attempted to assess the effects of daily supplementation with 10 μg vitamin D and 1,000 mg calcium, consumed over an average intervention period of 7 years (habitual dietary intake not reported). Results were provided for the whole study group as well as for those that were not using personal supplements at baseline. The study found no significant effect of the intervention on overall total fracture risk.

SACN (2016) identified one RCT already considered by the IOM and that used a single high annual dose of vitamin D (Sanders et al., 2010), reported mixed evidence from three meta-analyses on vitamin D supplementation and fracture prevention, and concluded that RCTs do not show an effect of vitamin D supplements on fracture risk in older men and women (≥ 50 years). One meta-analysis of 19 RCTs (12 on vitamin D, 7 on a vitamin D metabolite, assessed separately) was supportive of a beneficial effect of vitamin D supplementation (D₂ or D₃, with or without calcium) of doses above 10 μg/day in reducing the risk of non-vertebral fractures (9 RCTs) and hip fractures (5 RCTs) (Bischoff-Ferrari et al., 2009b). In contrast, the two other meta-analyses (of 53 and 22 RCTs, respectively) showed that 'vitamin D' alone had no effect on fracture risk (Avenell et al., 2014; Bolland et al., 2014b). Avenell et al. (2014) did not exclude studies using supplementation with vitamin D metabolites and only Bischoff-Ferrari et al. (2009b) included exclusively studies based on oral vitamin D supplementation (12 on oral vitamin D₂ or D₃, out of 19 RCTs included). All three systematic reviews included studies on institutionalised subjects; few included studies were published in 2010 or afterwards, i.e. after the IOM report; and several studies were in common in these three reviews. The systematic review by Avenell et al. (2014) included a subgroup analysis on community-dwelling subjects receiving vitamin D plus calcium (versus placebo or no treatment), indicating that hip fracture incidence was not significantly reduced with vitamin D supplementation of 10–20 μg/day. A similar finding on the lack of effect on the risk of total or hip fracture of vitamin D plus calcium in community-dwelling subjects was reported by Bolland et al. (2014b). The Panel considers that no conclusion can be drawn from these systematic reviews for the setting of DRVs for vitamin D.

For the present Section, the Panel considered a population-based Swedish cohort, which included 61,433 women (born between 1917 and 1948, mean age between 56 and 59 years of subjects in quintiles of vitamin D intake) followed for 19 years (Snellman et al., 2014). Total dietary intakes (from foods and supplements) were assessed repeatedly by several FFQs. Women with a total intake higher than 12.5 μg/day did not have a lower rate of fracture of any type, compared with those with a total vitamin D intake below 3.5 μg/day. Calcium intake (higher or less than 800 mg/day) did not modify these results. The Panel notes that, in this study, dietary intakes of vitamin D, from foods and supplements, was not associated with the rate of fractures in community-dwelling middle-aged women.

The Panel notes that the available evidence does not indicate that vitamin D supplementation up to 20 μg/day has a significant positive effect on fracture risk in community-dwelling adults with adequate calcium intakes.

5.2.3 Muscle strength/function and physical performance in adults

IOM ( 2011 ) (Sections 4, 5.1.2.1.4 and Appendix B) noted that randomised trials suggest that vitamin D dosages of at least 20 μg/day, with or without calcium, may improve physical performance measures, but that the evidence was insufficient to define the shape of the dose–response curve. The findings by Lamberg-Allardt et al. (2013), Newberry et al. (2014) and SACN (2016) have been described previously (Section 5.1.2.1.4).

The Panel takes into account the same nine RCTs with heterogeneous designs, which were considered in relation to associations between serum 25(OH)D concentrations and muscle strength/function and physical performance (Section 5.1.2.1.4). From these, only three provided data on habitual dietary intake of vitamin D: means at baseline of 1.6 and 4.1 μg/day in the placebo and intervention groups, respectively (Pirotta et al., 2015); median habitual dietary intake of vitamin D of about 4.3–4.8 μg/day (Wood et al., 2014); means of 10.4 and 10.5 μg/day at baseline and end of study (Uusi-Rasi et al., 2015).

Overall, these nine RCTs do not provide evidence for an effect of vitamin D supplementation (10 to about 71 μg/day), with or without calcium, on these outcomes. However, one study showed a beneficial effect of vitamin D supplementation (vs placebo) on postural stability in the subgroup of subjects with elevated baseline body sway (Lips et al., 2010). Another one showed a beneficial effect of vitamin D supplementation with calcium (vs calcium alone) on muscle strength and mobility in those who were the weakest and slowest at baseline (Zhu et al., 2010). A third one found a beneficial effect of vitamin D supplementation (two different doses compared) on the ability to do chair-stand tests in subjects with the slowest gait speed at baseline (Lagari et al., 2013). These three studies used doses ranging between 10 and 50 μg/day.

The Panel notes that these studies suggest that vitamin D supplementation does not generally affect muscle strength/function and indices of physical performance. However, subgroup analyses on small numbers of older subjects, with impaired indices of physical performance at baseline, indicated beneficial effects of vitamin D supplementation doses (ranging between 10 and 50 μg/day) in three of these studies.

5.2.4 Risk of falls and falling in adults

IOM ( 2011 ) (Sections 4, 5.1.2.1.5 and Appendix B) concluded, based on Cranney et al. (2007) and Chung et al. (2009) and an additional literature search, that, some RCTs found a significant effect of vitamin D supplementation on fall incidence or risk or number of fallers, but the greater part of the 20 RCTs considered found no effect of supplemental vitamin D (usually with doses of 10–20 μg/day and 50 μg/day in one), with or without supplemental calcium, on the risk of falls. A number of RCTs analysed falls rather than fallers.

Newberry et al. (2014) identified two RCTs that examined the effect of supplementation with vitamin D and calcium on the risk of falls/falling among community-dwelling older adults (Prince et al., 2008; Pfeifer et al., 2009) considered by IOM (2011). Prince et al. (2008) supplemented older women daily with 25 μg vitamin D₂ and 1,000 mg calcium or only 1,000 mg calcium in a 1-year RCT and found a significantly decreased risk of falling at least once, and a decreased risk for first falls, especially in winter/spring, in the group that received vitamin D₂. In the 1-year RCT performed by Pfeifer et al. (2009), older individuals received daily either 20 μg vitamin D₃ and 1,000 mg calcium or only 1,000 mg calcium and found a reduction in the number of first fallers in the group that received vitamin D₃.

The Panel notes the above-mentioned RCTs in populations with different mean baseline serum 25(OH)D concentration (Sections 5.1.2.1.4, 5.1.2.1.5 and 5.2.3). The RCT by Wood et al. (2014) showed no effect of vitamin D₃ supplementation (10 or 25 μg/day versus placebo) on the number of 'ever fallen' falls in healthy post-menopausal women. The RCT by Uusi-Rasi et al. (2015), that investigated the effect of daily vitamin D₃ supplementation (20 μg) with or without exercise (versus placebo with or without exercise) on falls in women at risk of falling, showed no effect of vitamin D₃ on the rate or risk of falls or injurious falls.

The Panel considers that, among studies identified by IOM (2011) and Newberry et al. (2014), some provide evidence of an effect on falls or the number of fallers with daily 20–25 μg vitamin D₂/D₃ with calcium in comparison with calcium alone in community-dwelling older adults, whereas two RCTs retrieved by the Panel thereafter in healthy post-menopausal women did not find such effect of vitamin D₃ compared with placebo.

5.2.5 Bone mineral density/content in infants and children

For infants, IOM ( 2011 ) identified two RCTs (Greer et al., 1982; Greer and Marshall, 1989), using supplemental doses of 10 μg/day vitamin D, and which found inconsistent effects on BMC (Sections 4, 5.1.2.2.1 and Appendix B).

The Panel takes into account the same two randomised trials (Holmlund-Suila et al., 2012; Gallo et al., 2013) that were considered in relation to associations between serum 25(OH)D concentrations and BMD/BMC (Section 5.1.2.2.1), and that used various doses of vitamin D₃ supplementation, without a placebo group, in (mostly) breastfed infants. Only one provided data on the vitamin D intake through breast milk between ages 1 and 12 months (1–6 μg/day) (Gallo et al., 2013). They showed that a supplementation with 10 μg/day vitamin D₃ was sufficient to reach a plasma/serum 25(OH)D of at least 50 nmol/L in (almost) all subjects, and that there was no significant differences in several bone measurements between groups.

For children, IOM ( 2011 ) considered five RCTs (Ala-Houhala et al., 1988b; Cheng et al., 2005; El-Hajj Fuleihan et al., 2006; Viljakainen et al., 2006a; Andersen et al., 2008) performed in children of various ages and receiving doses of vitamin D between 5 and about 50 μg/day (Sections 4, 5.1.2.2.1 and Appendix B). Only three of them provided data on habitual dietary intake of vitamin D. Three studies did not find an effect of these doses on BMC/BMD, while one study found an effect with 5 and 10 μg/day only in subjects with compliance above 80% (but not in the ITT analysis) and another with 50 μg/day.

The Panel takes into account the same RCT that was considered in relation to associations between serum 25(OH)D concentrations and BMD/BMC (Section 5.1.2.2.1). Molgaard et al. (2010) supplemented 12-year-old girls with either placebo, 5 or 10 μg vitamin D/day for 1 year, in addition to the habitual dietary intake of vitamin D (mean intakes of 2.6, 2.8 and 2.5 μg/day, respectively) and found no effect on BMC/BMD.

The Panel notes that the data available on vitamin D supplementation in infants (10 μg/day or higher) and children (5–50 μg/day) and BMD/BMC are inconsistent. The Panel, however, notes that two recent trials showed that a supplementation with 10 μg/day vitamin D₃ in (mostly) breastfed infants was sufficient to reach a plasma/serum 25(OH)D of at least 50 nmol/L in (almost) all subjects.

5.2.6 Pregnancy, lactation and related outcomes in mothers and infants

For pregnancy, IOM ( 2011 ) (Sections 4, 5.1.3, 5.1.4 and Appendix B) considered one RCT that found no effect of maternal vitamin D supplementation in combination with calcium on the incidence of preeclampsia (Marya et al., 1987), and reported on four RCTs that found no effect of maternal vitamin D supplementation on birth weight or length of the children (Brooke et al., 1980; Maxwell et al., 1981; Mallet et al., 1986; Marya et al., 1988). In these studies, the supplementation was generally based on doses of 25–30 μg/day, and started at various time points in pregnancy.

SACN (2016) identified one RCT (Cooper et al., 2016) (Section 5.1.3), which did not show any significant difference in neonatal whole body BMC (primary outcome) between intervention group (maternal supplementation of 25 μg/day vitamin D₃ from about 14–17 weeks of gestation and until delivery) and placebo (EFSA, 2016).

The Panel takes into account the same paper by Wagner et al. (2013b) that was considered in relation to associations between serum 25(OH)D concentrations and health outcomes in pregnancy (Section 5.1.3). This paper reported on pooled data from two RCTs in which daily supplementation doses of 50 and 100 μg vitamin D₃ during pregnancy had no effect on neonatal birth weight, and risk of pre-eclampsia or preterm birth in pregnant women with mean serum 25(OH)D concentrations of 57–65 nmol/L at baseline. The Panel did not retrieve any relevant RCT on vitamin D intake/supplementation during lactation and relevant outcomes in mother or child.

The Panel notes that the number of RCTs that focused on effects of supplementation during pregnancy or lactation on outcomes related to, e.g. bone, pre-eclampsia and birth weight, is small. The doses used in the few studies reported varies between 25 and 100 μg/day, with no effect on the variables studied. In addition, the amount of vitamin D in human milk is modestly correlated with maternal vitamin D intake (unless 'high' supplemental doses are used, Section 2.3.7.2).

5.2.7 Overall conclusions on vitamin D intake from supplements and musculoskeletal health outcomes, pregnancy and lactation, in relation to the setting of DRVs for vitamin D

The Panel concludes that:

there is some evidence suggesting that beneficial effects of vitamin D supplementation on BMD/BMC may be achieved with doses of about 10 to 25 μg/day in free-living subjects with 25(OH)D concentrations between 29 and 50 nmol/L, and that the effects may depend on calcium intake,
available studies suggest that vitamin D supplementation does not generally affect muscle strength/function and indices of physical performance. However, subgroup analyses on small numbers of older subjects, with impaired indices of physical performance at baseline, indicated beneficial effects of vitamin D supplementation doses (ranging between 10 and 50 μg/day) in three studies,
although results of available studies on vitamin D supplementation with or without calcium are not entirely consistent, there is some evidence for an effect on the risk of falls/falling with daily 20–25 μg vitamin D supplementation with calcium in comparison with calcium alone, in community-dwelling older subjects,
available studies provide no evidence for an effect of vitamin D supplementation on fracture risk,
the available data do not allow conclusion to be drawn on an effect of vitamin D supplementation on BMD/BMC in infants and children. However, two recent trials showed that a supplementation with 10 μg/day vitamin D₃ in (mostly) breastfed infants was sufficient to reach a plasma/serum 25(OH)D of at least 50 nmol/L in (almost) all subjects,
available studies provide no evidence for an effect of vitamin D supplementation on a number of outcomes in pregnancy or lactation.

Overall, the Panel notes that there may be beneficial effect of vitamin D supplementation above 10 μg/day (in addition to the habitual dietary intake of vitamin D) on some musculoskeletal health outcomes, particularly in subjects with compromised musculoskeletal health or 'low' 25(OH)D concentration. Habitual dietary intake of vitamin D is generally low (Section 3); however, the Panel notes that, in these supplementation studies with heterogeneous designs, vitamin D intake from foods was reported only in a limited number of trials. In addition, the extent to which cutaneous vitamin D synthesis has contributed to the vitamin D supply, and thus may have confounded the relationship between vitamin D intake and reported outcomes, is not known. The Panel concludes that these data are not useful as such for setting DRVs for vitamin D. For the purpose of deriving DRVs for vitamin D, these data may only be used to support the outcome of the characterisation of the vitamin D intake-status relationship undertaken by the Panel under conditions of assumed minimal endogenous vitamin D synthesis (Section 5.3).

5.3 Vitamin D intake and serum 25(OH)D concentration

The relationship between vitamin D intake and serum 25(OH)D concentrations has been investigated in numerous intervention studies in all age groups including different doses of vitamin D provided as supplements or as foods or fortified foods.

The systematic reviews by Cranney et al. (2007) and Chung et al. (2009), which were used by IOM ( 2011 ), included RCTs using supplements or fortified foods. Focusing on 28 RCTs (26 on adults), Chung et al. (2009) concluded that a relationship between increasing supplementation doses of vitamin D₃ and increasing net change in serum 25(OH)D concentration was evident in both adults and children, that the dose–response relationships differed depending on serum 25(OH)D concentration of the participants at baseline (< 40 nmol/L vs > 40 nmol/L), and depending on the duration of supplementation (< 3 months vs > 3 months). The range of supplementation doses was large (5–125 μg/day), the baseline serum 25(OH)D concentrations varied and the assays used for measuring serum 25(OH)D concentrations were heterogeneous. Supplementation with vitamin D₂ was more commonly used than supplementation with vitamin D₃ in RCTs in infants and pregnant or lactating women, with a resulting significant increase in serum 25(OH)D concentrations in infants or lactating mothers and in cord blood. Based on Cranney et al. (2007) and Chung et al. (2009) and some new RCTs, IOM (2011) undertook specific meta-regression analyses to obtain a dose–response curve, in order to set DRVs for vitamin D (Section 5.3.1).

Lamberg-Allardt et al. (2013) considered the results from four systematic reviews (Cranney et al., 2007; Chung et al., 2009; Cashman et al., 2011a; Black et al., 2012) (Section 5.3.1 for Cashman et al. (2011a)) on the relationship between vitamin D supplementation/fortification and serum 25(OH)D concentrations, and underlined the important issue of the heterogeneity in the results according to the assays used to measure serum 25(OH)D concentrations (Section 2.4.1). Lamberg-Allardt et al. (2013) concluded that the systematic reviews indicated a clear effect of supplementation and fortified foods on the serum 25(OH)D concentration, but the doses needed to achieve specific concentrations of 25(OH)D are difficult to determine. One systematic review (Black et al., 2012) estimated that 1 μg vitamin D ingested only from fortified foods increased the serum 25(OH)D concentration by 1.2 nmol/L (heterogeneity index (I²) = 89%, adjusted R² = 0.67). Habitual dietary intake of vitamin D was usually not reported in the 16 RCTs included in this review thus was not added to the content of the fortified foods for the data analysis.

Newberry et al. (2014) identified one systematic review (Autier et al., 2012) that included 76 placebo-controlled and open-label trials published from 1984 through 2011 and addressed the relationship between supplementation with vitamin D₂ or D₃ (oral or injection, with or without calcium, with vitamin D doses ranging from 5 to 250 μg/day (median: 20 μg/day)) and net change in serum 25(OH)D concentrations. The meta-regression analysis by Autier et al. (2012) of serum 25(OH)D concentration on (log-transformed) vitamin D doses (< 100 μg/day) showed that serum 25(OH)D concentrations increased by an average of 1.95 nmol/L for each 1 μg per day vitamin D₃ supplementation (without calcium). In this analysis, vitamin D₂ supplementation resulted in smaller increases compared with vitamin D₃ supplementation, and simultaneous supplementation with calcium resulted in non-significantly smaller increases in serum 25(OH)D concentrations. As the number of trials that used higher doses of vitamin D was small (n = 3 with doses of 100 μg/day or more), whether the dose–response relationship reaches a plateau at higher doses could not be assessed. Newberry et al. (2014) noted that most studies included in Autier et al. (2012) did not stratify findings by sex, and the review itself did not stratify findings by assay method. In addition to the systematic review by Autier et al. (2012), Newberry et al. (2014) identified eighteen new RCTs (in addition to those included by Chung et al. (2009)) (two of them using fortified foods, the others using vitamin D supplements with or without calcium, one study using a vitamin D₂ supplement). Overall, all studies reported an increase in serum 25(OH)D with vitamin D supplementation. Newberry et al. (2014) also provided plots showing the relationship between vitamin D₃ supplementation doses and net changes in serum 25(OH)D concentrations in 44 RCTs, according to populations (adults and children), baseline serum 25(OH)D concentrations, duration of supplementation, and assay used to assess serum 25(OH)D concentration.

The Panel notes that studies based on vitamin D supplementation and/or food and food fortification suggest a relationship between vitamin D intake and serum 25(OH)D concentrations in all ages and that this relationship depends on several factors, including baseline serum 25(OH)D concentrations, supplementation dose, study duration, and assay used to assess serum 25(OH)D concentration.

5.3.1 Characterisation of the intake-status relationship in previous approaches

One approach to assess the intake-status relationship could be to rely on a sample of individual data from a particular study (e.g. regression analysis on individual data). The Panel did not have access to a sufficiently large and representative sample of individual data from a study considered relevant for the aim of setting DRVs at the European level.

Several bodies have characterised the intake-status relationship through meta-regression approaches, which has also been the target of various authors (e.g. Cashman et al. (2011a); Autier et al. (2012)). In a meta-regression approach, a quantitative synthesis of the dose–response relationship between mean results at group level from studies is usually carried out (taking into account potential confounders by relevant adjustments). Once the methodological heterogeneity is characterised, the remaining variation reflects a real phenomenon that describes the extent to which different populations behave differently. One advantage of the meta-regression approach is the representativity, by considering several studies from various populations in different contexts, instead of relying on specific data from one specific study undertaken in a particular context. However, by using group means from studies, the information on the variability between individuals is diminished, which may complicate the setting of, e.g. a reference value that would correspond to the intake needed to cover the requirements of 97.5% of individuals. The CI in meta-regression analyses provides an estimate of the uncertainty about the fitted response line due to sampling, but does not provide an estimate of the variability between individuals (Section 5.3.2).

IOM (2011) carried out meta-regression analyses of the relationship between serum 25(OH)D concentrations and log-transformed (ln) total intake of vitamin D (from food and supplements) during winter at latitudes above 49.5°N in Europe or Antarctica, separately for children/adolescents, young/middle-aged adults, and older adults (Ala-Houhala et al., 1988b; Van Der Klis et al., 1996; Schou et al., 2003; Larsen et al., 2004; Viljakainen et al., 2006a, 2009; Cashman et al., 2008, 2009; Smith et al., 2009).27 The IOM considered that the response of serum 25(OH)D concentration to vitamin D intake is non-linear, the rise being steeper below 25 μg/day and flattening above 25 μg/day. Baseline serum 25(OH)D concentrations and age did not have a significant effect in the response of serum 25(OH)D concentration to total vitamin D intake. The IOM performed also a meta-regression analysis on all age-groups (6 to more than 60 years) at latitudes above 49.5°N using the CI around the mean. The IOM performed as well a separate analysis for latitudes 40–49°N during winter. In particular, this analysis (i) showed that the achieved serum 25(OH)D concentration at these lower latitudes was greater (24%) for a given total intake compared to that achieved in the previous analysis at higher latitudes, and (ii) explained less variability than the model at higher latitudes. Thus, the IOM decided to focus on latitude above 49.5°N to set DRVs for vitamin D. The IOM noted that, at a total intake of 10 μg/day, the predicted mean serum 25(OH)D concentration was 59 nmol/L in children and adolescents, young and middle-aged adults, and older adults (with a lower limit of the CI of about 52 nmol/L). The IOM also noted that, at a total intake of 15 μg/day, the predicted mean serum 25(OH)D concentration was 63 nmol/L (lower limit of the CI of 56 nmol/L). These results were used to set the EAR and RDA for vitamin D, which take into account the uncertainties in these analyses (Section 4).

Cashman et al. (2011a) applied a meta-regression approach using different model constructs (curvilinear as in the approach by the IOM, or linear) to explore the most appropriate model of the relationship between total vitamin D intake (from food and supplements) and serum 25(OH)D concentration. Priority was given to data from winter-based RCTs performed at latitudes 49.5–78°N, using vitamin D₃ supplementation (not vitamin D₂) in children and adults (i.e. excluding infants, pregnant and lactating women) and with a duration of at least 6 weeks (Harris and Dawson-Hughes, 2002). Thus, 12 RCTs in 11 references were included (Ala-Houhala et al., 1988b; Honkanen et al., 1990; Pfeifer et al., 2001; Meier et al., 2004; Barnes et al., 2006; Viljakainen et al., 2006b, 2009; Cashman et al., 2008, 2009, 2011b; Smith et al., 2009). When the included RCTs did not assess and/or did not report the habitual vitamin D intake (Ala-Houhala et al., 1988b; Honkanen et al., 1990; Pfeifer et al., 2001; Meier et al., 2004), the authors considered the mean intake of the relevant age and sex group, from the national nutrition survey preferably from the country in which the RCT was performed. A combined weighted linear model meta-regression analysis of log-transformed (ln) total vitamin D intake (maximum 50 μg/day) versus achieved serum 25(OH)D in winter produced a curvilinear relationship. Use of non-transformed total vitamin D intake data (maximum 35 μg/day, Section 2.4.1 and Aloia et al. (2008)) provided a linear relationship. At an intake of 15 μg/day (i.e. the RDA set by the IOM for vitamin D for adults aged 19–70 years, Section 4), the predicted serum 25(OH)D concentration at the lower limit of the 95% CI of the log-transformed and the linear models was 54.4 and 55.2 nmol/L, respectively. The total vitamin D intake estimated to achieve the 'RDA-type' and 'EAR-type' values for 25(OH)D concentrations set by the IOM (50 and 40 nmol/L, Section 4) was 9 μg/day for 50 nmol/L (and 2.7 μg/day for 40 nmol/L) in the log-transformed model. In the linear model, this intake was 12 μg/day for 50 nmol/L (and 6.5 μg/day for 40 nmol/L), respectively. In further publications of the same author, use of a 95% prediction interval (PI) in meta-regression analyses was considered to allow for estimation of the requirement of 97.5% of the population (Cashman and Kiely, 2014; Cashman, 2015).

The Nordic Council of Ministers (2014) performed two meta-regression analyses of log₁₀ (serum 25(OH)D) versus total vitamin D intake. The included studies were selected mainly from the systematic review by Cashman et al. (2011a) and the previous Nordic recommendations (NNR, 2004), and studies using doses of vitamin D higher than 30 μg/day were excluded. The first meta-regression analysis included six supplementation studies pertinent to the Nordic countries, undertaken in adults (≤ 60 years) (Barnes et al., 2006; Cashman et al., 2008; Viljakainen et al., 2009) and children (Ala-Houhala et al., 1988b; Molgaard et al., 2010; Cashman et al., 2011b), during winter, at latitudes 50–61°N. The response to intake was found to be limited or absent for baseline concentrations above 50 nmol/L. It was considered that an intake of 7.2 μg/day would maintain a mean serum concentration during winter of about 50 nmol/L for 50% of subjects. Using the lower limit of the 95% CI, it was considered that about 10 μg/day would be sufficient for most of the population. The second meta-regression analysis was based on supplementation studies in mainly older adults (> 65 years) (Sem et al., 1987; Pfeifer et al., 2001; Meier et al., 2004; Viljakainen et al., 2006b; Cashman et al., 2009) during winter at latitudes 51–61°N. It was considered that an intake of about 5 μg/day would maintain a mean serum 25(OH)D concentration of about 50 nmol/L during wintertime. This estimate was lower than for younger adults, but the 95% CI was wider and, based on its lower bound, it was considered that an intake of about 10–11 μg/day is sufficient for most of this population. These results were used to set the reference values for vitamin D in the Nordic Countries (Section 4).

The Panel applied the meta-regression approach to assess the intake-status relationship with the aim to set DRVs for vitamin D.

5.3.2 Characterisation of the intake-status relationship by EFSA in adults and children

As indicated previously (Section 2.3.1), the Panel considered that the association between vitamin D intake and status for the purpose of deriving DRVs for vitamin D should be assessed under conditions of minimal endogenous vitamin D synthesis.

5.3.2.1 Methods

As preparatory work for the setting of DRVs for vitamin D, a comprehensive literature search and review was performed to identify and summarise studies that could be used to assess the dose–response relationship between vitamin D₂ or vitamin D₃ intake (i.e. oral exposure) and plasma/serum 25(OH)D concentration (Brouwer-Brolsma et al., 2016).

Prospective studies (that primarily aimed to investigate the dose–response association of vitamin D intake and status) and trials that investigated vitamin D intake and 25(OH)D concentration, published through July 2014 were systematically searched and reviewed (Brouwer-Brolsma et al., 2016). Studies were eligible for inclusion if they:

were conducted in humans of all ages,
investigated oral exposure to vitamin D₂ or vitamin D₃ at least twice a week via diet, supplements or fortified foods and its subsequent effect/association on/with 25(OH)D concentration,
were performed in a period of assumed minimal endogenous vitamin D synthesis, i.e. at latitudes above 40°N from October through April (or below 40°S from April through October).28 Additional further selections were also proposed (Brouwer-Brolsma et al., 2016), based on the UV-index (UV-index < 3) or a simulation model (Webb, 2006; Webb and Engelsen, 2006) (Section 2.3.1), but in the end were not applied in this analysis, as it would have led to a substantial reduction in the number of arms (53% and 86% of the 83 arms would have been excluded, respectively),
and lasted for at least 6 weeks (Sections 2.4.1 and 5.3.1).

More information on the inclusion/exclusion criteria and the selection process can be found in Brouwer-Brolsma et al. (2016).

Finally, 56 articles matched the eligibility criteria, reporting on data of 65 relevant studies (e.g. one article reporting data in children and in adults was considered as one article reporting data on two studies). The majority of the included studies were trials (n = 57), investigating the effects of supplements, fortified foods or foods naturally rich in vitamin D (fish). Only eight prospective cohort studies fulfilled the inclusion criteria.

Using a meta-analytic approach, EFSA performed quantitative syntheses of the summary data extracted by Brouwer-Brolsma et al. (2016) from the included studies (Appendix C). Data from prospective observational studies identified were analysed but were not included in the meta-regression dose–response model by EFSA, which was based solely on randomised trials data.

The 57 trials included in the preparatory literature review represented 141 arms. Of these 141 arms, EFSA excluded a total of 58 arms from the analysis (25 arms excluded from the preliminary data set of 116 arms, and 33 arms further excluded from the final data set of 83 arms, Table 8, Appendix D.A), in particular:

arms from trials on population groups other than children and adults (i.e. infants, pregnant women, lactating women, as these populations represent particular age and/or physiological conditions and the number of arms were low29),
arms resulting in total intakes exceeding the UL set for adults (EFSA NDA Panel, 2012a) (Section 2.2.2.2),
arms in which vitamin D₂ was administered. In view of the conflicting results regarding the potential differences in the biological potencies and catabolism of vitamin D₂ and D₃ (Sections 2.3.2 and 2.3.6), and the low number of arms using vitamin D₂ (six), this exclusion was considered appropriate by the Panel.
arms for which methodological and/or statistical inconsistencies were identified.

This left 83 arms from 35 trials in the analysis (Appendix D.B), of which nine arms were on children (age range: 2–17 years) and the other arms were on adults (mean age between 22 and 86 years).

The continuous outcome, i.e. plasma/serum 25(OH)D concentration, was analysed by EFSA using the summary data extracted for each arm in each individual study. Background intake was added by EFSA to the supplemental vitamin D dose to generate total vitamin D intake estimates. If the habitual vitamin D intake of the cohort(s) within a study was not reported in the papers, surrogates were imputed using the appropriate age- and sex- specific mean vitamin D intake values (from food) from the national nutrition survey relevant to the country in which the study was performed (17 trials) (AppendixC and Table 11of Appendix D.B).

Two different models of the dose–response relationship between total vitamin D intake and plasma/serum 25(OH)D concentration were explored (Appendix C): a linear model and a non-linear model (i.e. with the natural logarithm transformation of the total intake). Finally, the Panel decided to retain the non-linear model to better describe the dose–response shape and to be able to include results from trials using higher supplemental doses (i.e. up to 50 μg/day).

A number of factors potentially influencing the dose–response relationship (Section 2) were investigated, in order to select factors to be included in the final model to characterise the high heterogeneity of results across individual trials. These were: total vitamin D intake, baseline serum concentration, study duration (≤ 3 months versus > 3 months; or ≤ 3 months, versus 3–6 months versus 1–2 years), latitude (as different categories), assay method (HPLC and LC–MS versus immunoassays; or each analytical method as an individual category), BMI (Section 2.3.5), co-supplementation with calcium, funding source, age, sex, risk of bias (RoB), assessment of compliance, period of study publication (before/after 2000) or study start period (as a 'proxy' to the temporal trends in assay method use, Section 2.4.1), and ethnicity (as a 'proxy' for skin pigmentation, possible effects of genotypes and some lifestyle habits (Sections 2.3.1 and 2.5) that were usually not reported in the included trials). In particular for ethnicity, the data were missing for almost half of the studies, as this information was not reported in the papers (Appendix C).

5.3.2.2 Results

The meta-regression analysis carried out on the selected arms resulted in two predictive equations of achieved serum 25(OH)D concentration:

y = 23.2 ln (total vitamin D intake in μg/day) (equation 1, unadjusted model)

and

y = 16.3 ln (total vitamin D intake) + 0.5 mean baseline 25(OH)D − 0.5 latitude + 0.9 study start year − 2.0 HPLC − 4.7 LC–MS + 0.6 CPBA − 6.4 ELISA/nr + 1.3 Other assay + 7.8 compliance not assessed (equation 2, adjusted model)

The model corresponding to equation 2 was adjusted for baseline concentration (continuous), latitude (continuous), study start year (continuous), type of analytical method applied (RIA as 'reference' category for the model, HPLC, LC-MS, CPBA, ELISA/not reported (nr), other30), assessment of compliance (yes as 'reference' category for the model, no/unknown). No interaction terms were introduced.

The 95% CI around the coefficient mentioned above for each variable are given in Table 5, Section 8.8 of Appendix C (e.g. about 14.4–18.2 for the coefficient of about 16.3 obtained for ln (total vitamin D intake)). The summary data of the included studies are given in Appendix D.B, in particular the mean and SD baseline and achieved serum 25(OH)D concentrations per included arm are given in Table 11 of this Appendix.

After the inclusion of the final set of covariates, the adjusted R ² (proportion of between-study variance explained) of the final model was 85%, meaning that the fitted factors were able to characterise most of the across-trials variability in response.

The two equations above were used to predict the achieved mean serum 25(OH)D concentrations corresponding to total vitamin D intakes of 5, 10, 15, 20, 50, 100 μg/day (Appendix C, Table 6) and to estimate the total vitamin D intakes that would achieve serum 25(OH)D concentrations of 50, 40, 30, 25 nmol/L (Appendix C, Table 7).

In the adjusted multivariable models, all covariates were set to their mean values: mean baseline serum 25(OH)D concentration: 50.7 nmol/L; mean latitude: 53°N; study start year: 2005; assay – HPLC: 10%; LC–MS: 18%; CPBA: 13%; ELISA: 20%; Other: 8%; compliance not assessed/unknown: 27%. As such the adjusted model predictions can be interpreted as referring to an average ideal population in which the major factors influencing the heterogeneity across different populations have been ruled out. Such a reduction in heterogeneity is reflected in the narrower PI as compared to the unadjusted model.

Lower and upper limits of the 95% CI and of the 95% PI were calculated for both the adjusted and the unadjusted model. In the meta-regression context, where a random-effects approach is applied:

the CI illustrates the uncertainty about the position of the regression line (i.e. across-study conditional means);
the PI illustrates the uncertainty about the true mean effect that would be predicted in a future study.

The 95% PI refers to the population of mean responses (not individual responses) as analysed in the random-effects model, while DRVs are set for healthy individuals and populations, if possible taking into account the distribution of requirements in the population (EFSA NDA Panel, 2010). As such, it is possible to think of the 95% PI of the meta-regression only as an approximation of the interval that would allow for estimation of the requirements for 95% of individuals in the overall population.

The role of BMI (Section 2.3.5) was tested and it was not included in the final model as a covariate, as it did not reach statistical significance and in consideration of potential ecological fallacy (Appendices C and D.C). Sex and age were also not included in the final model, as they did not further explain between-study variability once mutually adjusted for all other factors (Appendices C and D.C). However, regarding the role of age, a stratified analysis was carried out (Tables 14 and 15 of Appendix D.G), to quantify the impact of the exclusions of the four trials on children (age range: 2–17 years, nine arms) on the predicted achieved mean serum 25(OH)D concentrations (Appendix C, Table 6) and estimated total vitamin D intakes (Appendix C, Table 7).

In the restricted data set of 74 arms on adults only, there was an overall small decrease in all serum estimates (and consequently a small increase in total intakes that would achieve target values). Overall estimates (Tables 14 and 15 of Appendix D.G) did not substantially change as compared to the full data set including children (Tables 6 and 7 of Appendix C). Thus, the Panel decided to retain the data on children and on adults in the dose–response analysis (Section6).
Children tended to achieve the same mean serum 25(OH)D concentrations as adults at a lower total intake (Table 14 of Appendix D.G). It was not possible to apply a full adjustment to estimate the values based only on the four children trials, as it would have required a much higher minimum number of 'points' per covariate (at least 10 arms for each included factor). Instead, values from a model adjusted for mean baseline 25(OH)D concentration were provided. As such these estimates are not directly comparable to the ones in the adjusted model in adults, as they are not adjusted for the same set of covariates. The unadjusted model showed lower average intakes, but estimates were less precise; also the highest dose investigated in the included arms was 10 μg/day, so predictions at higher intakes are extrapolations from the model. For these reasons results from the models on children data could only be evaluated qualitatively.

A number of sensitivity analyses were also carried out by EFSA to evaluate whether the findings were robust to the assumptions made in the systematic review protocol and the analyses (Appendix C), in particular, on the background intake imputation process, on eligibility criteria (e.g. fortified food trials versus supplement trials, cf. Section 2.3.2); characteristics of participants (e.g. exclusion of trials that did not explicitly exclude supplement users, persons on specific medications, persons using sunbeds/artificial UV-B sources or going on sunny holidays, Table 16 of Appendix D.H). None of these sensitivity analyses raised serious concerns about the robustness of the overall analysis. In addition, there was no particular indication of publication bias as explored on the subset of trials for which the mean difference in response could be estimated (Appendices C and D.J).

The Panel considers that the results of this meta-regression analysis can be used to set DRVs for vitamin D. The meta-regression model of serum 25(OH)D response to ln of total vitamin D intake from the adjusted model (n = 83 arms) is shown in Figure 3, as well as in Appendix D.F (for comparison with the unadjusted model).

Meta-regression model of serum 25(OH)D response to ln of total vitamin D intake (adjusted model) (n = 83 arms)

Circles represent mean achieved 25(OH)D in all included arms, either control or intervention arms. Mean achieved values are modelled against total vitamin D intake and adjusted in the multivariable approach by mean baseline values, latitude, study start year, analytical method for measuring serum 25(OH)D and compliance (Appendix C). The confidence interval (CI) illustrates the uncertainty about the position of the regression line (blue). The prediction interval (PI) illustrates the uncertainty about the true mean effect that would be predicted in a future study.

5.3.3 Qualitative overview of available data on infants, children, pregnant or lactating women

Only two studies (Ala-Houhala et al., 1986; Atas et al., 2013) that were conducted in breastfed infants met the eligibility criteria of the comprehensive literature search (Brouwer-Brolsma et al., 2016) mentioned previously (Section 5.3.2.1) (in situation of assumed minimal endogenous vitamin D synthesis). Both studies included an intervention group that was allocated to 10 μg/day vitamin D. Atas et al. (2013) also included a study group that was allocated to 5 μg/day. Ala-Houhala et al. (1986) supplemented with vitamin D₂ from birth for the duration of 15 weeks. At baseline, mean serum 25(OH)D concentrations were approximately 20 nmol/L, which rose to roughly 80 nmol/L after 15 weeks (values estimated from figures). Atas et al. (2013) supplemented with vitamin D₃ from birth for the duration of 17 weeks, but did not measure baseline serum 25(OH)D concentration. Follow-up measurements at 4 months of age showed, however, higher serum 25(OH)D concentrations than in the study by Ala-Houhala et al. (1986): serum 25(OH)D reached a median (min–max) concentration of 99 (43–265) nmol/L in the 5 μg group, and 141 (80–375) nmol/L in the 10 μg group (Atas et al., 2013).

Three prospective studies (Sullivan et al., 2005; Lehtonen-Veromaa et al., 2008; Andersen et al., 2013) met the eligibility criteria of the comprehensive literature search (Brouwer-Brolsma et al., 2016) mentioned previously (Section 5.3.2.1). Two of these studies reported on dietary vitamin D intake (food only; Sullivan et al., 2005; Lehtonen-Veromaa et al., 2008); one study measured vitamin D intake covering both dietary as well as supplemental intake (Andersen et al., 2013). Vitamin D intakes ranged from median (IQR) 3.9 (1.9–7.0) μg/day ((Andersen et al., 2013), dietary and supplemental intake) to mean of 5.4 ± 1.4 ((Sullivan et al., 2005), dietary intake only). Follow-up time ranged from one (Andersen et al., 2013) to 4 years (Lehtonen-Veromaa et al., 2008). Mean (± SD) age at baseline ranged from 11 ± 1 (Sullivan et al., 2005) to 16 ± 2 (Lehtonen-Veromaa et al., 2008) years old. All three studies performed the baseline and follow-up 25(OH)D measurements in February/March. In one study (Andersen et al., 2013), baseline vitamin D intake was (median (IQR)) 3.9 (1.9–7.0) μg/day, food and supplements) and serum 25(OH)D concentrations at follow-up were (median (IQR)) 23(17–36) nmol/L. For the two others (Sullivan et al., 2005; Lehtonen-Veromaa et al., 2008), baseline vitamin D intakes (food only) were (mean ± SD) 5.4 ± 1.4 and 4.0 ± 2.2 μg/day, while serum 25(OH)D concentrations at follow-up were 50 ± 14 and 48 ± 17 nmol/L, respectively.

Two RCTs on pregnant or lactating women met the eligibility criteria of the comprehensive literature search (Brouwer-Brolsma et al., 2016) mentioned previously (Section 5.3.2.1).

In an open-label trial, Ala-Houhala et al. (1986) examined the effect of vitamin D supplementation on serum 25(OH)D concentration in pregnant women (41 starters, 39 completers) living in Finland (61°N), delivering in January, and whose age was not reported. Eight women were supplemented with 12.5 μg vitamin D₃ per day throughout the pregnancy; 33 others did not receive any supplementation.31 Background dietary vitamin D and calcium intakes were not assessed. Serum 25(OH)D was measured only at the delivery (thus at the end of the supplementation period). At delivery, there was a pronounced difference in mean (± SEM) serum 25(OH)D concentrations between women that received vitamin D supplementation (57 ± 11 nmol/L) and those that did not (25 ± 2 nmol/L) (p < 0.01).

In the same open-label trial, Ala-Houhala et al. (1986) also studied the effect of vitamin D supplementation in lactating women (49 starters, 49 completers)32 (whose age was not reported) from January through March. Mothers received either no treatment (n = 16), 25 μg (n = 16) or 50 μg (n = 17) vitamin D₃ per day from delivery until 15 weeks post partum. Background dietary vitamin D and calcium intakes were not assessed. At baseline, there were no significant differences in serum 25(OH)D concentrations across the three groups, showing mean concentrations around 32 nmol/L (concentration is estimated from figure in original article). However, after 15 weeks, serum 25(OH)D concentration significantly increased in the treatment groups (p < 0.01). That is, up to about 75 nmol/L in the 25 μg/day group and 100 nmol/L in the 50 μg/day group (concentrations are estimated from figure in original article).

The Panel considers that the two infant studies may be used to set DRVs for vitamin D in infants (Section 6.2), while the other available studies on children, and pregnant or lactating women are not informative for the setting of DRVs for vitamin D for these population groups. The Panel also notes that the European Society for Paediatric Gastroenterology, Hepatology, and Nutrition (ESPGHAN) (Braegger et al., 2013) recommends the 'pragmatic use' of a serum 25(OH)D concentration of > 50 nmol/L to indicate sufficiency and a daily supplement of 10 μg to all infants. The Panel notes that mean vitamin D concentrations in breast milk of healthy lactating women, unsupplemented or supplemented with vitamin D, are low (0.25–2.0 μg/L) (Section 2.3.7.2), that maternal vitamin D intake during lactation influences maternal serum 25(OH)D concentration, but is only modestly correlated with the amount of vitamin D in human milk, unless high supplemental doses are used. Thus, the Panel considers that the derivation of a DRV for infants in the second half of the first year of life by extrapolation from the vitamin D intake of exclusively breastfed infants (in the first half year of life) is not possible, and that the derivation of DRVs for vitamin D for lactating women based on the compensation of the vitamin D loss in breast milk is not justified.

6 Data on which to base dietary reference values

In spite of the high variability in serum 25(OH)D measurements obtained with different analytical methods, the Panel concludes that serum 25(OH)D concentration, which reflects the amount of vitamin D attained from both cutaneous synthesis and dietary sources, can be used as a biomarker of vitamin D status in adult and children populations. Serum 25(OH)D concentration can also be used as a biomarker of vitamin D intake in a population with low exposure to UV-B irradiation.

The Panel considers some musculoskeletal health outcomes as suitable to set DRVs for vitamin D for (healthy) adults, infants and children (Sections 5.1.1–5.1.5). Taking into account the overall evidence and uncertainties on the relationship between serum 25(OH)D concentration and the risk of these health outcomes, the Panel concludes that a serum 25(OH)D concentration of 50 nmol/L is a suitable target value for all age and sex groups (Section 5.1.6). For setting DRVs for vitamin D, the Panel considers the dietary intake of vitamin D necessary to achieve this serum 25(OH)D concentration. As for other nutrients, DRVs for vitamin D are set assuming that intakes of interacting nutrients, such as calcium (EFSA NDA Panel, 2015b) (Section 2.3.8), are adequate.

The Panel considers that the available evidence (Sections 5.1.6, 5.2.7 and 5.3.2) does not allow the setting of ARs and PRIs and chooses to set AIs instead, for all population groups.

6.1 Adults

The Panel used information obtained from characterising the intake-status relationship for vitamin D (meta-regression described in Section 5.3.2 and related Appendices C and D) to derive the vitamin D intake to achieve a target serum 25(OH)D concentration of 50 nmol/L. For the purpose of deriving AIs for vitamin D, the Panel decided to focus on the adjusted model of achieved mean serum 25(OH)D according to ln (total vitamin D intake) (i.e. total intake from habitual diet, fortified foods or supplements). As indicated in Section 5.3.2, this adjusted model was obtained with data mostly on adults (74 arms out of 83 included arms) in randomised trials using vitamin D ₃ (not vitamin D₂) (Sections 2.3.2, 2.3.6, 5.3.2 and Appendix C), and the estimates from this adjusted model were derived based on all covariates set to their mean values.

In the adjusted model, the total intake estimated to achieve a serum 25(OH)D concentration of 50 nmol/L, as identified by the lower limit of the 95% PI, is 16.1 μg/day (Appendix C, Table 7). Equally, at a vitamin D intake of 15 μg/day, the predicted mean serum 25(OH)D concentration is 63 nmol/L (95% CI: 58–69 nmol/L), with a predicted value at the lower limit of the 95% PI of 49 nmol/L (Appendix C, Table 6).

The Panel supports the use of the prediction interval (and not only the confidence interval) in the context of the meta-regression analysis and notes that the PI in the context of a meta-regression analysis illustrates the uncertainty about the true mean response predicted in a future study (Section 5.3). The Panel also considers that the 95% PI of the mean responses can provide useful indications on the interval that would include true individual responses from the populations of interest (PI of the individual responses). The extent to which it can cover the 95% PI of individual responses can vary quite significantly depending on specific conditions and assumptions, such as the distribution that can be assumed for the biomarker within each study and how much the available studies are representative of the target population. In the current case, an empirical analysis has shown that the use of the 95% PI lower limit of the mean response (serum 25(OH)D) concentration) can support the statement that the majority of the adult population will achieve the target serum concentration.

The Panel therefore sets an AI for vitamin D for adults at 15 μg/day, considering that, at this intake, the majority of the adult population will achieve the target serum 25(OH)D concentration near or above 50 nmol/L. The Panel notes that this value for total intake of vitamin D is above the supplementation dose identified in Section 5.2.7 in relation to beneficial effect on musculoskeletal health outcomes. The Panel decided not to set specific AIs for 'younger' or 'older' adults, because there was no evidence of a significant difference in absorption capacity of vitamin D between 'younger' and 'older' healthy adults (Section 2.3.2) and the majority of the studies used to set the target value for 25(OH)D concentration were carried out in 'older' free-living healthy adults (Section 5.1).

The unadjusted model (Appendix C, Table 6) can be also taken into account as it encompasses the whole heterogeneity across trials. In this unadjusted model, considering a vitamin D intake of 15 μg/day, (i) the lower limit of the 95% PI is 34 nmol/L and (ii) the upper limit of the 95% PI is 91 nmol/L (while 78 nmol/L in the adjusted model). The Panel notes that the value of 34 nmol/L is above the concentrations that have been observed in relation to overt adverse health outcomes (Sections 5.1.2.1.2 and 5.1.2.1.6) and that the value of 91 nmol/L (or 78 nmol/L) is in the physiological range.

The Panel underlines that the meta-regression was done on data collected under conditions of assumed minimal cutaneous vitamin D synthesis. In the presence of endogenous cutaneous vitamin D synthesis (Section 2.3.1), the requirement for dietary vitamin D is lower or may even be zero.

6.2 Infants aged 7–11 months

The Panel notes that there are few data on the relationship between serum 25(OH)D concentration and the risk of adverse musculoskeletal health outcomes available in infants (Section 5.1.2.2). The Panel notes that there are no data to suggest a different target value for serum 25(OH)D concentration for infants compared to the adult age group (Section 5.1.6). The Panel also notes that there is a general agreement that human milk does not contain sufficient vitamin D to prevent rickets in the breastfed infant (Olafsdottir et al., 2001), that a vitamin D intake of 10 μg/day was considered adequate for the majority of infants in the first half year of life (EFSA NDA Panel, 2013) and that ESPGHAN Committee on Nutrition recommends a daily oral supplementation of 10 μg vitamin D for all infants during the first year of life starting from birth onwards (Braegger et al., 2013). The Panel considers that, since breast milk does not supply adequate amounts of vitamin D to the breastfed infant (Section 2.3.7.2), the derivation of an AI for infants in the second half of the first year of life by extrapolation from the vitamin D intake of exclusively breastfed infants (in the first half year) is not possible (Section 5.3.3).

In line with conclusions by the IOM (Section 4), the Panel notes that two recent trials (Holmlund-Suila et al., 2012; Gallo et al., 2013) (Sections 5.1.2.2 and 5.2.6) showed that a supplementation with 10 μg/day vitamin D₃ in (mostly) breastfed infants was sufficient to reach a plasma/serum 25(OH)D of at least 50 nmol/L in (almost) all subjects. Only two studies (Ala-Houhala et al., 1986; Atas et al., 2013) that were conducted in breastfed infants in situation of assumed minimal endogenous vitamin D synthesis met the eligibility criteria of the comprehensive literature search (Brouwer-Brolsma et al., 2016) mentioned previously (Sections 5.3.2 and 5.3.3). Giving vitamin D supplementation of 10 μg/day to breastfed infants for at least 15 weeks led to an achieved mean/median serum 25(OH)D concentration of at least 80 nmol/L in both studies.

The Panel sets an AI for vitamin D for infants aged 7–11 months at 10 μg/day.

6.3 Children

The Panel notes that there are few data on the relationship between serum 25(OH)D concentration and the risk of adverse musculoskeletal health outcomes available in children (Section 5.1.2.2). The Panel notes that there are no data to suggest a different target value for serum 25(OH)D concentration for children compared to the adult age group (Section 5.1.6).

The Panel sets an AI for vitamin D for adults at 15 μg/day, based on the analysis of the adjusted and unadjusted models of the meta-regression analysis (Sections 5.3.2 and 6.1, and Appendix C) that were obtained from data collected mostly on adults, but also on children. Thus, this value of 15μg/day may also apply to children.

From Appendices C and D.G, a further stratified analysis by age group (adults versus children) (Section 5.3.2) showed that children tended to achieve the same mean serum 25(OH)D concentrations as the adults at a lower total intake (Appendix D.G). In addition, in the analysis based only on the four trials in children (age range: 2–17 years, nine arms), taking into account the limitations previously described in details (Section 5.3.2):

In the adjusted model (adjusted only for baseline serum 25(OH)D concentration), the total intake estimated to achieve a serum 25(OH)D concentration of 50 nmol/L (Appendix D.G, Table 15), at the lower limit of the 95% CI, is 7.9 μg/day and at the lower limit of the 95% PI is 10.9 μg/day. In the unadjusted model, the total intake estimated to achieve a serum 25(OH)D concentration of 50 nmol/L, at the lower limit of the 95% CI, is 11.5 μg/day and, at the lower limit of the 95% PI, is 27.6 μg/day.
Equally, at a vitamin D intake of 15 μg/day (Appendix D.G, Table 14), in the adjusted model (adjusted only for baseline mean serum 25(OH)D), the predicted mean serum 25(OH)D concentration is 67 nmol/L (95% CI: 61–73 nmol/L), with a predicted value at the lower limit of the 95% PI of 55 nmol/L. In the unadjusted model, at a vitamin D intake of 15 μg/day, the predicted mean serum 25(OH)D concentration is 73 nmol/L (95% CI: 56–91 nmol/L), with a predicted value at the lower limit of the 95% PI of 35 nmol/L. The Panel notes that this value of 35 nmol/L is above the concentrations that have been observed in relation to overt adverse health outcomes (Section 5.1.2.2.2).

The Panel sets an AI for vitamin D for all children (1–17 years) at 15 μg/day. The Panel underlines that the meta-regression was done on data collected under conditions of assumed minimal cutaneous vitamin D synthesis. In the presence of endogenous cutaneous vitamin D synthesis (Section 2.3.1), the requirement for dietary vitamin D is lower or may even be zero.

6.4 Pregnancy

The Panel notes that there are no data to suggest a different target value for 25(OH)D concentration for pregnant women compared to non-pregnant women (Sections 5.1.3 and 5.1.6).

The Panel considers that the AI for pregnant women is the same as for non-pregnant women (15 μg/day). The Panel underlines that the meta-regression on adults (Sections 5.3.2 and 6.1) was done on data collected under conditions of assumed minimal cutaneous vitamin D synthesis. In the presence of endogenous cutaneous vitamin D synthesis (Section 2.3.1), the requirement for dietary vitamin D is lower or may even be zero.

6.5 Lactation

The Panel notes that no studies were available for setting an AI for lactating women (Sections 5.1.4 and 5.1.6). The Panel notes that mean vitamin D concentrations in breast milk of healthy lactating women, unsupplemented or supplemented with vitamin D, are low (0.25–2.0 μg/L), that maternal vitamin D intake during lactation influences maternal serum 25(OH)D concentration, but is only modestly correlated with the amount of vitamin D in human milk, unless high supplemental doses are used. The Panel considers that the derivation of DRVs for vitamin D for lactating women based on the compensation of the vitamin D loss in breast milk is not justified (Sections 2.3.7 and 5.3.3).

The Panel considers that the AI for lactating women is the same as for non-lactating women (15 μg/day). The Panel underlines that the meta-regression on adults (Sections 5.3.2 and 6.1) was done on data collected under conditions of assumed minimal cutaneous vitamin D synthesis. In the presence of endogenous cutaneous vitamin D synthesis (Section 2.3.1), the requirement for dietary vitamin D is lower or may even be zero.

Conclusions

The Panel concludes that ARs and PRIs for vitamin D cannot be derived for adults, infants and children, and therefore defines AIs, for all population groups. The Panel considers that serum 25(OH)D concentration, which reflects the amount of vitamin D attained from both cutaneous synthesis and dietary sources, can be used as a biomarker of vitamin D intake in adult and children populations with low exposure to UV-B irradiation and as a biomarker of vitamin D status. The Panel notes that the evidence on the relationship between serum 25(OH)D concentration and the risk of musculoskeletal health outcomes in (healthy) adults, infants and children, and some adverse pregnancy-related health outcomes, is widely variable. Several factors contribute to this, and also include the large variation in the results from different laboratories and assays used for measuring serum 25(OH)D concentrations. Taking into account the overall evidence and uncertainties, the Panel considers that a serum 25(OH)D concentration of 50 nmol/L is a suitable target value for all population groups, in view of setting the AIs for vitamin D.

For adults, the Panel sets an AI for vitamin D at 15 μg/day. This is based on an adjusted model of the meta-regression analysis of serum 25(OH)D concentration according to total vitamin D intake (natural log of the sum of habitual diet, and fortified foods or supplements using vitamin D₃). The Panel considers that, at this intake, the majority of the adult population will achieve a serum 25(OH)D concentration near or above the target of 50 nmol/L. For children aged 1–17 years, the Panel sets an AI for vitamin D at 15 μg/day, based on the meta-regression analysis. For infants aged 7–11 months, the Panel sets an AI for vitamin D at 10 μg/day, based on four recent trials on the effect of vitamin D supplementation on serum 25(OH)D concentration in (mostly) breastfed infants. For pregnant and lactating women, the Panel considers that the AI is the same as for non-pregnant non-lactating women, i.e. 15 μg/day. The Panel underlines that the meta-regression in adults and children was done on data collected under conditions of assumed minimal cutaneous vitamin D synthesis. In the presence of endogenous cutaneous vitamin D synthesis, the requirement for dietary vitamin D is lower or may even be zero (Table 4).

Table 4. Summary of dietary reference values for vitamin D

Age	AIa ^a Under conditions of assumed minimal cutaneous vitamin D synthesis. In the presence of endogenous cutaneous vitamin D synthesis (Section 2.3.1), the requirement for dietary vitamin D is lower or may be even zero. (μg/day)
7–11 months	10
1–3 years	15a ^a Under conditions of assumed minimal cutaneous vitamin D synthesis. In the presence of endogenous cutaneous vitamin D synthesis (Section 2.3.1), the requirement for dietary vitamin D is lower or may be even zero.
4–6 years	15a ^a Under conditions of assumed minimal cutaneous vitamin D synthesis. In the presence of endogenous cutaneous vitamin D synthesis (Section 2.3.1), the requirement for dietary vitamin D is lower or may be even zero.
7–10 years	15a ^a Under conditions of assumed minimal cutaneous vitamin D synthesis. In the presence of endogenous cutaneous vitamin D synthesis (Section 2.3.1), the requirement for dietary vitamin D is lower or may be even zero.
11–14 years	15a ^a Under conditions of assumed minimal cutaneous vitamin D synthesis. In the presence of endogenous cutaneous vitamin D synthesis (Section 2.3.1), the requirement for dietary vitamin D is lower or may be even zero.
15–17 years	15a ^a Under conditions of assumed minimal cutaneous vitamin D synthesis. In the presence of endogenous cutaneous vitamin D synthesis (Section 2.3.1), the requirement for dietary vitamin D is lower or may be even zero.
≥ 18 yearsb ^b Including pregnancy and lactation.	15a ^a Under conditions of assumed minimal cutaneous vitamin D synthesis. In the presence of endogenous cutaneous vitamin D synthesis (Section 2.3.1), the requirement for dietary vitamin D is lower or may be even zero.

^a Under conditions of assumed minimal cutaneous vitamin D synthesis. In the presence of endogenous cutaneous vitamin D synthesis (Section 2.3.1), the requirement for dietary vitamin D is lower or may be even zero.
^b Including pregnancy and lactation.

Recommendations for Research

There is a need for further research to study the respective impact of vitamin D dietary intake and sunlight exposure on serum 25(OH)D concentrations. Future studies should investigate food-based strategies to ensure adequate vitamin D intakes accounting for latitude, sunlight exposure and diet.

Studies are needed that are specifically designed to identify cut-off values for serum 25(OH)D concentration or other suitable biomarkers for vitamin D status to derive DRVs for vitamin D for infants, children, adults, pregnant and lactating women.

Standardised investigations are needed to assess changes in musculoskeletal health outcomes (and surrogate markers) in response to vitamin D₂ and D₃ intake, and in relation to serum 25(OH)D concentrations.

The potential mechanisms of the cause and effect relationships between vitamin D and non-musculoskeletal health outcomes should be further explored.

There is a need for studies that assess the different diets of infants, in particular those consuming infant or follow-on formulas and processed cereal-based foods fortified with vitamin D in addition to vitamin D supplements.

More data on the effects of genotype and body fat mass on vitamin D metabolism and the requirements for vitamin D are warranted.

More precise data on total vitamin D concentration in foods would also be useful. Studies investigating the effect of 25(OH)D naturally occurring in foods on serum 25(OH)D concentration are also suggested.

Notes

1 Scientific Committee for Food, 1993. Nutrient and energy intakes for the European Community. Reports of the Scientific Committee for Food, 31^st series. Food – Science and Technique, European Commission, Luxembourg, 248 pp.

2 Regulation (EC) No 178/2002 of the European Parliament and of the Council of 28 January 2002 laying down the general principles and requirements of food law, establishing the European Food Safety Authority and laying down procedures in matters of food safety. OJ L 31, 1.2.2002, p. 1–24.

3 For conversion between µg and International Units (IU) of vitamin D intake: 1 µg = 40 IU and 0.025 µg = 1 IU. For conversion between nmol/L and ng/mL for serum 25(OH)D concentration: 2.5 nmol/L = 1 ng/mL.

4 http://www.who.int/uv/intersunprogramme/activities/uv_index/en/index3.html

5 Vitamin D antirachitic activity in milk was assessed through measurement of vitamin D₂, vitamin D₃, 25(OH)D₂, and 25(OH)D₃ concentrations in the milk and conversion of findings into biological activity values with reference data from biological activity assays.

6 http://www.deqas.org/

7 http://www.nist.gov/mml/csd/organic/vitamindinserum.cfm

8 https://ods.od.nih.gov/Research/vdsp.aspx

9 Regulation (EC) No 1925/2006 of the European Parliament and of the Council of 20 December 2006 on the addition of vitamins and minerals and of certain other substances to foods. OJ L 404, 30.12.2006, p. 26.

10 Directive 2002/46/EC of the European Parliament and of the Council of 10 June 2002 on the approximation of the laws of the Member States relating to food supplements. OJ L 183, 12.7.2002, p. 51.

11 Commission Directive 2006/141/EC of 22 December 2006 on infant formulae and follow-on formulae and amending Directive 1999/21/EC. OJ L 401, 30.12.2006, p. 1. and Commission Directive 2006/125/EC of 5 December 2006 on processed cereal-based foods and baby foods for infants and young children. OJ L 339, 06.12.2006, p. 16.

12 Further abbreviated into NCM in tables.

13 No EAR or lower RNI were set.

14 Defined as a value 'judged to be a level or range of intake at which there is no risk of deficiency, and below a level where there is a risk of undesirable effects' (DH, 1991).

15 Personal communication from one author: vitamin D₃.

16 Age, race, clinic, season of blood draw, physical activity, weight, and height.

17 Per cent of body fat, or health status, or neuromuscular measures (unable to complete chair stand or narrow walk, grip strength), or hip BMD, or falls.

18 Fracture risk in the second quartile was not statistically different from the one in fourth quartile.

19 Adjusted for age, country of birth, BMI, physical activity, season of blood draw, previous low-trauma fracture after age 50 years, calcium supplement, and vitamin D supplement.

20 Additional adjustments for falls or BMD or neuromuscular measures (chair stands and narrow walk test) or serum 1,25(OH)₂D or multivariate model excluding subjects taking vitamin D supplements.

21 Some studies also on vitamin D metabolites/analogues were considered in these systematic reviews.

22 Some of these studies are described below. Others were undertaken, e.g. with vitamin D metabolite or based on a frequency of supplementation (e.g. once per 3 months) that did not match the inclusion criteria of the Panel (Section 5.1).

23 The SPPB includes an assessment of standing balance, gait speed (4-m walking speed) and 5-time chair stand tests.

24 The Modified Cooper test is used as a measurement of overall mobility.

25 Leg muscle quality (LMQ) defined as the level of force produced per unit of muscle mass.

26 None of the included studies in this systematic review were published in 2010 or afterwards.

27 All these studies used vitamin D₃ supplementation.

28 Based on the protocol by Brouwer-Brolsma et al. (2016).

29 Two arms on pregnant women, three arms on lactating women, three arms on infants.

30 Based on the data reported by the contractor. 'Other' covers methods presented as 'enzyme immunoassays', Nichols method, 'chemoluminescence immunoassays', 'immunoenzymetric assay' in the references included by the contractor.

31 The study by Ala-Houhala et al. (1986) also included a third study group, including women that were supplemented during the second trimester of the pregnancy. As 25(OH)D measurements were only conducted at delivery, the data of this group that was supplemented in the second trimester were not considered relevant to the review by Brouwer-Brolsma et al. (2016). (i.e. supplementation was terminated several months before the 25(OH)D measurements were conducted).

32 Researchers already followed these lactating women during pregnancy, during which women were distributed over three groups: i.e. eight women were supplemented with 12.5 µg vitamin D₃ per day throughout the pregnancy; eight women were supplemented with 12.5 µg vitamin D₃ per day during the second trimester of pregnancy; 33 others did not receive any supplementation. After delivery, the women were redistributed into three 'new' groups, as explained in the paragraph above.

33 Vitamin D₂ or D₃ +/− calcium compared with calcium or placebo, vitamin D₃ compared with placebo, vitamin D₃ + calcium compared with calcium.

34 Vitamin D₃ compared with placebo, vitamin D₃ + calcium compared placebo.

35 Vitamin D₃ compared with placebo, vitamin D₃ + calcium compared with calcium, vitamin D₃ + calcium compared placebo.

36 Vitamin D₂ or D₃ +/− calcium compared with calcium or placebo.

37 In four trials using vitamin D₃ with end of study 25(OH)D concentrations of > 74 nmol/L, out of 10 trials reporting follow-up or change in mean 25(OH)D concentrations.

38 Older adults receiving vitamin D₂ or D₃ with calcium, compared to calcium or placebo (three trials on total fractures), or vitamin D₃ with calcium, compared to placebo (two trials on hip fractures, combined OR: 0.69; 95 % CI: 0.53–0.90).

39 In a sensitivity analysis, Cranney et al. (2007) found that combining the results from eight trials on oral vitamin D₂ or D₃ with calcium, compared to placebo or calcium alone, showed a significant reduction in the risk of falls (OR: 0.84 ; 95% CI: 0.76–0.93), heterogeneity I² = 0%).

40 In total, Cranney et al. (2007) identified one RCT, three cohort studies and one case–control study on the association between serum 25(OH)D concentrations and risk of falls, as well as three RCTs and four cohort studies on the association between 25(OH)D concentrations and measures of performance (among these, one cohort investigated both risk of falls and measures of performance). Chung et al. (2009) identified three additional RCTs on vitamin D supplementation and the risk of falls, including one which also investigated measures of performance, and one additional RCTs on vitamin D with calcium and measures of performance.

41 Indicated as 'first author date a' or 'first author date b' or 'first author date c' in case two (or three) different populations were included in the same study, e.g. normal weight, overweight and obese people.

42 Barger-Lux et al., 1998; DeLappe et al., 2006; Goussous et al., 2005; Pekkarinen et al., 2010; and Vieth et al., 2001.

43 Only one trial (four arms) was undertaken in the Southern hemisphere (at 78°S) (Smith et al., 2009). All the other trials included were undertaken in the Northern hemisphere (41°N – 63°N) (Table 9 in Appendix D.B).

References

Abrams SA, Hicks PD and Hawthorne KM, 2009. Higher serum 25-hydroxyvitamin D levels in school-age children are inconsistently associated with increased calcium absorption. Journal of Clinical Endocrinology and Metabolism, 94, 2421– 2427.
Abrams SA, Hawthorne KM, Rogers SP, Hicks PD and Carpenter TO, 2012. Effects of ethnicity and vitamin D supplementation on vitamin D status and changes in bone mineral content in infants. BioMed Central Pediatrics, 12, 6.
Abrams SA, Hawthorne KM and Chen Z, 2013. Supplementation with 1000 IU vitamin D/d leads to parathyroid hormone suppression, but not increased fractional calcium absorption, in 4-8-y-old children: a double-blind randomized controlled trial. American Journal of Clinical Nutrition, 97, 217– 223.
Afssa (Agence française de sécurité sanitaire des aliments), 2001. Apports nutritionnels conseillés pour la population française. Editions Tec & Doc, Paris, France. 605 pp.
Ahn J, Yu K, Stolzenberg-Solomon R, Simon KC, McCullough ML, Gallicchio L, Jacobs EJ, Ascherio A, Helzlsouer K, Jacobs KB, Li Q, Weinstein SJ, Purdue M, Virtamo J, Horst R, Wheeler W, Chanock S, Hunter DJ, Hayes RB, Kraft P and Albanes D, 2010. Genome-wide association study of circulating vitamin D levels. Human Molecular Genetics, 19, 2739– 2745.
Ajibade DV, Benn BS and Christakos S, 2010. Mechanism of action of 1,25-dihydroxyvitamin D3 on intestinal calcium absorption and renal calcium transport. In: MF Holick (ed). Vitamin D, Physiology, Molecular Biology, and Clinical Applications. Humana Press, New York, NY, USA. pp. 175– 188.
Akcakus M, Koklu E, Budak N, Kula M, Kurtoglu S and Koklu S, 2006. The relationship between birthweight, 25-hydroxyvitamin D concentrations and bone mineral status in neonates. Annals of Tropical Paediatrics, 26, 267– 275.
Aksnes L and Aarskog D, 1982. Plasma concentrations of vitamin D metabolites in puberty: effect of sexual maturation and implications for growth. Journal of Clinical Endocrinology and Metabolism, 55, 94– 101.
Al-oanzi ZH, Tuck SP, Raj N, Harrop JS, Summers GD, Cook DB, Francis RM and Datta HK, 2006. Assessment of vitamin D status in male osteoporosis. Clinical Chemistry, 52, 248– 254.
Ala-Houhala M, 1985. 25-Hydroxyvitamin D levels during breast-feeding with or without maternal or infantile supplementation of vitamin D. Journal of Pediatric Gastroenterology and Nutrition, 4, 220– 226.
Ala-Houhala M, Koskinen T, Terho A, Koivula T and Visakorpi J, 1986. Maternal compared with infant vitamin D supplementation. Archives of Disease in Childhood, 61, 1159– 1163.
Ala-Houhala M, Koskinen T, Parviainen MT and Visakorpi JK, 1988a. 25-Hydroxyvitamin D and vitamin D in human milk: effects of supplementation and season. American Journal of Clinical Nutrition, 48, 1057– 1060.
Ala-Houhala M, Koskinen T, Koskinen M and Visakorpi JK, 1988b. Double blind study on the need for vitamin D supplementation in prepubertal children. Acta Paediatrica Scandinavica, 77, 89– 93.
Alfaham M, Woodhead S, Pask G and Davies D, 1995. Vitamin D deficiency: a concern in pregnant Asian women. British Journal of Nutrition, 73, 881– 887.
Aloia JF, Talwar SA, Pollack S and Yeh J, 2005. A randomized controlled trial of vitamin D3 supplementation in African American women. Archives of Internal Medicine, 165, 1618– 1623.
Aloia JF, Patel M, Dimaano R, Li-Ng M, Talwar SA, Mikhail M, Pollack S and Yeh JK, 2008. Vitamin D intake to attain a desired serum 25-hydroxyvitamin D concentration. American Journal of Clinical Nutrition, 87, 1952– 1958.
Aloia JF, Chen DG, Yeh JK and Chen H, 2010. Serum vitamin D metabolites and intestinal calcium absorption efficiency in women. American Journal of Clinical Nutrition, 92, 835– 840.
Aloia JF, Dhaliwal R, Shieh A, Mikhail M, Fazzari M, Ragolia L and Abrams SA, 2014. Vitamin D supplementation increases calcium absorption without a threshold effect. American Journal of Clinical Nutrition, 99, 624– 631.
Andersen R, Molgaard C, Skovgaard LT, Brot C, Cashman KD, Jakobsen J, Lamberg-Allardt C and Ovesen L, 2008. Effect of vitamin D supplementation on bone and vitamin D status among Pakistani immigrants in Denmark: a randomised double-blinded placebo-controlled intervention study. British Journal of Nutrition, 100, 197– 207.
Andersen R, Brot C, Jakobsen J, Mejborn H, Molgaard C, Skovgaard LT, Trolle E, Tetens I and Ovesen L, 2013. Seasonal changes in vitamin D status among Danish adolescent girls and elderly women: the influence of sun exposure and vitamin D intake. European Journal of Clinical Nutrition, 67, 270– 274.
Anderson FH, Smith HE, Raphael HM, Crozier SR and Cooper C, 2004. Effect of annual intramuscular vitamin D-3 supplementation on fracture risk in 9440 community-living older people: The Wessex fracture prevention trial. Journal of Bone and Mineral Research, 19, S57.
Anses/CIQUAL (Agence nationale de sécurité sanitaire de l'alimentation, de l'environnement et du travail/Centre d'information sur la qualité des aliments), 2012. French food composition table version 2012. Available online:http://www.afssa.fr/TableCIQUAL/index.htm
Arabi A, Baddoura R, El-Rassi R and El-Hajj Fuleihan G, 2012. PTH level but not 25 (OH) vitamin D level predicts bone loss rates in the elderly. Osteoporosis International, 23, 971– 980.
Ardawi MS, Nasrat HA and BA'Aqueel HS, 1997. Calcium-regulating hormones and parathyroid hormone-related peptide in normal human pregnancy and postpartum: a longitudinal study. European Journal of Endocrinology, 137, 402– 409.
Arnaud SB, Stickler GB and Haworth JC, 1976. Serum 25-hydroxyvitamin-D in infantile rickets. Pediatrics, 57, 221– 225.
Aspray TJ and Francis RM, 2013. What can we learn about vitamin D requirements from post-mortem data? Osteoporosis International, 24, 1769– 1770.
Atas E, Karademir F, Ersen A, Meral C, Aydinoz S, Suleymanoglu S, Gultepe M and Gocmen I, 2013. Comparison between daily supplementation doses of 200 versus 400 IU of vitamin D in infants. European Journal of Pediatrics, 172, 1039– 1042.
Autier P and Gandini S, 2007. Vitamin D supplementation and total mortality: a meta-analysis of randomized controlled trials. Archives of Internal Medicine, 167, 1730– 1737.
Autier P, Gandini S and Mullie P, 2012. A systematic review: influence of vitamin D supplementation on serum 25-hydroxyvitamin D concentration. Journal of Clinical Endocrinology and Metabolism, 97, 2606– 2613.
Autier P, Boniol M, Pizot C and Mullie P, 2014. Vitamin D status and ill health: a systematic review. Lancet Diabetes Endocrinology, 2, 76– 89.
Avenell A, Gillespie WJ, Gillespie LD and O'Connell D, 2009. Vitamin D and vitamin D analogues for preventing fractures associated with involutional and post-menopausal osteoporosis. Cochrane Database of Systematic Reviews, 2, CD000227.
Avenell A, Mak JC and O'Connell D, 2014. Vitamin D and vitamin D analogues for preventing fractures in post-menopausal women and older men. Cochrane Database of Systematic Reviews, 4, CD000227.
Azar M, Basu A, Jenkins AJ, Nankervis AJ, Hanssen KF, Scholz H, Henriksen T, Garg SK, Hammad SM, Scardo JA, Aston CE and Lyons TJ, 2011. Serum carotenoids and fat-soluble vitamins in women with type 1 diabetes and preeclampsia: a longitudinal study. Diabetes Care, 34, 1258– 1264.
Baeksgaard L, Andersen KP and Hyldstrup L, 1998. Calcium and vitamin D supplementation increases spinal BMD in healthy, postmenopausal women. Osteoporosis International, 8, 255– 260.
Baker AM, Haeri S, Camargo CA Jr, Espinola JA and Stuebe AM, 2010. A nested case-control study of midgestation vitamin D deficiency and risk of severe preeclampsia. Journal of Clinical Endocrinology and Metabolism, 95, 5105– 5109.
Baker AM, Haeri S, Camargo CA Jr, Stuebe AM and Boggess KA, 2011. A nested case-control study of first-trimester maternal vitamin D status and risk for spontaneous preterm birth. American Journal of Perinatology, 28, 667– 672.
Bakhtiyarova S, Lesnyak O, Kyznesova N, Blankenstein MA and Lips P, 2006. Vitamin D status among patients with hip fracture and elderly control subjects in Yekaterinburg, Russia. Osteoporosis International, 17, 441– 446.
Balasubramanian K, Rajeswari J, Gulab Govil YC, Agarwal AK, Kumar A and Bhatia V, 2003. Varying role of vitamin D deficiency in the etiology of rickets in young children vs. adolescents in northern India. Journal of Tropical Pediatrics, 49, 201– 206.
Barbour KE, Houston DK, Cummings SR, Boudreau R, Prasad T, Sheu Y, Bauer DC, Tooze JA, Kritchevsky SB, Tylavsky FA, Harris TB, Cauley JA and Health ABCS, 2012. Calciotropic hormones and the risk of hip and nonspine fractures in older adults: the Health ABC Study. Journal of Bone and Mineral Research, 27, 1177– 1185.
Barger-Lux MJ, Heaney RP, Dowell S, Chen TC and Holick MF, 1998. Vitamin D and its major metabolites: serum levels after graded oral dosing in healthy men. Osteoporosis International, 8, 222– 230.
Barnes MS, Robson PJ, Bonham MP, Strain JJ and Wallace JM, 2006. Effect of vitamin D supplementation on vitamin D status and bone turnover markers in young adults. European Journal of Clinical Nutrition, 60, 727– 733.
Baroncelli GI, 2008. Quantitative ultrasound methods to assess bone mineral status in children: technical characteristics, performance, and clinical application. Pediatric Research, 63, 220– 228.
Barr R, Macdonald H, Stewart A, McGuigan F, Rogers A, Eastell R, Felsenberg D, Gluer C, Roux C and Reid DM, 2010. Association between vitamin D receptor gene polymorphisms, falls, balance and muscle power: results from two independent studies (APOSS and OPUS). Osteoporosis International, 21, 457– 466.
Barrett-Connor E, Laughlin GA, Li H, Nielson CM, Wang PY, Dam TT, Cauley JA, Ensrud KE, Stefanick ML, Lau E, Hoffman AR and Orwoll ES and Osteoporotic Fractures in Men Research Group, 2012. The association of concurrent vitamin D and sex hormone deficiency with bone loss and fracture risk in older men: the osteoporotic fractures in men (MrOS) study. Journal of Bone and Mineral Research, 27, 2306– 2313.
Basile LA, Taylor SN, Wagner CL, Horst RL and Hollis BW, 2006. The effect of high-dose vitamin D supplementation on serum vitamin D levels and milk calcium concentration in lactating women and their infants. Breastfeeding Medicine, 1, 27– 35.
Beaudart C, Buckinx F, Rabenda V, Gillain S, Cavalier E, Slomian J, Petermans J, Reginster JY and Bruyere O, 2014. The effects of vitamin D on skeletal muscle strength, muscle mass, and muscle power: a systematic review and meta-analysis of randomized controlled trials. Journal of Clinical Endocrinology and Metabolism, 99, 4336– 4345.
Beck-Nielsen SS, Jensen TK, Gram J, Brixen K and Brock-Jacobsen B, 2009. Nutritional rickets in Denmark: a retrospective review of children's medical records from 1985 to 2005. European Journal of Pediatrics, 168, 941– 949.
Berry D and Hyppönen E, 2011. Determinants of vitamin D status: focus on genetic variations. Current Opinion in Nephrology and Hypertension, 20, 331– 336.
Bhimma R, Pettifor JM, Coovadia HM, Moodley M and Adhikari M, 1995. Rickets in black children beyond infancy in Natal. South African Medical Journal, 85, 668– 672.
Biancuzzo RM, Clarke N, Reitz RE, Travison TG and Holick MF, 2013. Serum concentrations of 1,25-dihydroxyvitamin D2 and 1,25-dihydroxyvitamin D3 in response to vitamin D2 and vitamin D3 supplementation. Journal of Clinical Endocrinology and Metabolism, 98, 973– 979.
Bienaime F, Prie D, Friedlander G and Souberbielle JC, 2011. Vitamin D metabolism and activity in the parathyroid gland. Molecular and Cellular Endocrinology, 347, 30– 41.
Bikle DD, Siiteri PK, Ryzen E and Haddad JG, 1985. Serum protein binding of 1,25-dihydroxyvitamin D: a reevaluation by direct measurement of free metabolite levels. Journal of Clinical Endocrinology and Metabolism, 61, 969– 975.
Bikle DD, Gee E, Halloran B, Kowalski MA, Ryzen E and Haddad JG, 1986. Assessment of the free fraction of 25-hydroxyvitamin D in serum and its regulation by albumin and the vitamin D-binding protein. Journal of Clinical Endocrinology and Metabolism, 63, 954– 959.
Binet A and Kooh SW, 1996. Persistence of Vitamin D-deficiency rickets in Toronto in the 1990s. Canadian Journal of Public Health, 87, 227– 230.
Binkley N and Lensmeyer G, 2010. 25-hydroxyvitamin D assays and their clinical utility. In: MF Holick (ed). Vitamin D, Physiology, Molecular Biology, and Clinical Applications. Humana Press, New York, NY, USA. pp. 383– 400.
Bischoff-Ferrari HA, 2010. Health effects of vitamin D. Dermatologic Therapy, 23, 23– 30.
Bischoff-Ferrari HA, Stahelin HB, Dick W, Akos R, Knecht M, Salis C, Nebiker M, Theiler R, Pfeifer M, Begerow B, Lew RA and Conzelmann M, 2003. Effects of vitamin D and calcium supplementation on falls: a randomized controlled trial. Journal of Bone and Mineral Research, 18, 343– 351.
Bischoff-Ferrari HA, Dietrich T, Orav EJ, Hu FB, Zhang Y, Karlson EW and Dawson-Hughes B, 2004. Higher 25-hydroxyvitamin D concentrations are associated with better lower-extremity function in both active and inactive persons aged > or =60 y. American Journal of Clinical Nutrition, 80, 752– 758.
Bischoff-Ferrari HA, Zhang Y, Kiel DP and Felson DT, 2005. Positive association between serum 25-hydroxyvitamin D level and bone density in osteoarthritis. Arthritis and Rheumatism, 53, 821– 826.
Bischoff-Ferrari HA, Orav EJ and Dawson-Hughes B, 2006. Effect of cholecalciferol plus calcium on falling in ambulatory older men and women - A 3-year randomized controlled trial. Archives of Internal Medicine, 166, 424– 430.
Bischoff-Ferrari HA, Dawson-Hughes B, Staehelin HB, Orav JE, Stuck AE, Theiler R, Wong JB, Egli A, Kiel DP and Henschkowski J, 2009a. Fall prevention with supplemental and active forms of vitamin D: a meta-analysis of randomised controlled trials. British Medical Journal, 339, b3692.
Bischoff-Ferrari HA, Willett WC, Wong JB, Stuck AE, Staehelin HB, Orav EJ, Thoma A, Kiel DP and Henschkowski J, 2009b. Prevention of nonvertebral fractures with oral vitamin D and dose dependency: a meta-analysis of randomized controlled trials. Archives of Internal Medicine, 169, 551– 561.
Bischoff-Ferrari HA, Dawson-Hughes B, Platz A, Orav EJ, Stahelin HB, Willett WC, Can U, Egli A, Mueller NJ, Looser S, Bretscher B, Minder E, Vergopoulos A and Theiler R, 2010. Effect of high-dosage cholecalciferol and extended physiotherapy on complications after hip fracture: a randomized controlled trial. Archives of Internal Medicine, 170, 813– 820.
Bischoff-Ferrari HA, Dawson-Hughes B, Orav EJ, Staehelin HB, Meyer OW, Theiler R, Dick W, Willett WC and Egli A, 2016. Monthly high-dose vitamin D treatment for the prevention of functional decline: a randomized clinical trial. JAMA Internal Medicine, 176, 175– 183.
Bjelakovic G, Gluud LL, Nikolova D, Whitfield K, Wetterslev J, Simonetti RG, Bjelakovic M and Gluud C, 2011. Vitamin D supplementation for prevention of mortality in adults. Cochrane Database of Systematic Reviews, 7, CD007470.
Bjorn Jensen C, Thorne-Lyman AL, Vadgard Hansen L, Strom M, Odgaard Nielsen N, Cohen A and Olsen SF, 2013. Development and validation of a vitamin D status prediction model in Danish pregnant women: a study of the Danish National Birth Cohort. PLoS ONE, 8, e53059.
Black LJ, Seamans KM, Cashman KD and Kiely M, 2012. An updated systematic review and meta-analysis of the efficacy of vitamin D food fortification. Journal of Nutrition, 142, 1102– 1108.
Bleicher K, Cumming RG, Naganathan V, Blyth FM, Le Couteur DG, Handelsman DJ, Waite LM and Seibel MJ, 2014. U-shaped association between serum 25-hydroxyvitamin D and fracture risk in older men: results from the prospective population-based CHAMP study. Journal of Bone and Mineral Research, 29, 2024– 2031.
Blum M, Dolnikowski G, Seyoum E, Harris SS, Booth SL, Peterson J, Saltzman E and Dawson-Hughes B, 2008. Vitamin D(3) in fat tissue. Endocrine, 33, 90– 94.
Bodnar LM, Catov JM, Simhan HN, Holick MF, Powers RW and Roberts JM, 2007. Maternal vitamin D deficiency increases the risk of preeclampsia. Journal of Clinical Endocrinology and Metabolism, 92, 3517– 3522.
Bodnar LM, Catov JM, Zmuda JM, Cooper ME, Parrott MS, Roberts JM, Marazita ML and Simhan HN, 2010. Maternal serum 25-hydroxyvitamin D concentrations are associated with small-for-gestational age births in white women. Journal of Nutrition, 140, 999– 1006.
Bodnar LM, Rouse DJ, Momirova V, Peaceman AM, Sciscione A, Spong CY, Varner MW, Malone FD, Iams JD, Mercer BM, Thorp JM Jr, Sorokin Y, Carpenter MW, Lo J, Ramin SM, Harper M, Eunice Kennedy Shriver National Institute of Child H and Human Development Maternal-Fetal Medicine Units N, 2013. Maternal 25-hydroxyvitamin d and preterm birth in twin gestations. Obstetrics and Gynecology, 122, 91– 98.
de Boer IH, Levin G, Robinson-Cohen C, Biggs ML, Hoofnagle AN, Siscovick DS and Kestenbaum B, 2012. Serum 25-hydroxyvitamin D concentration and risk for major clinical disease events in a community-based population of older adults: a cohort study. Annals of Internal Medicine, 156, 627– 634.
Bolland MJ, Bacon CJ, Horne AM, Mason BH, Ames RW, Wang TK, Grey AB, Gamble GD and Reid IR, 2010. Vitamin D insufficiency and health outcomes over 5 y in older women. American Journal of Clinical Nutrition, 91, 82– 89.
Bolland MJ, Grey A, Gamble GD and Reid IR, 2014a. Vitamin D supplementation and falls: a trial sequential meta-analysis. Lancet Diabetes Endocrinology, 2, 573– 580.
Bolland MJ, Grey A, Gamble GD and Reid IR, 2014b. The effect of vitamin D supplementation on skeletal, vascular, or cancer outcomes: a trial sequential meta-analysis. Lancet Diabetes Endocrinology, 2, 307– 320.
Bolton-Smith C, McMurdo ME, Paterson CR, Mole PA, Harvey JM, Fenton ST, Prynne CJ, Mishra GD and Shearer MJ, 2007. Two-year randomized controlled trial of vitamin K1 (phylloquinone) and vitamin D3 plus calcium on the bone health of older women. Journal of Bone and Mineral Research, 22, 509– 519.
Bonjour JP, Benoit V, Payen F and Kraenzlin M, 2013. Consumption of yogurts fortified in vitamin D and calcium reduces serum parathyroid hormone and markers of bone resorption: a double-blind randomized controlled trial in institutionalized elderly women. Journal of Clinical Endocrinology and Metabolism, 98, 2915– 2921.
Bonjour JP, Kohrt W, Levasseur R, Warren M, Whiting S and Kraenzlin M, 2014. Biochemical markers for assessment of calcium economy and bone metabolism: application in clinical trials from pharmaceutical agents to nutritional products. Nutrition Research Reviews, 27, 252– 267.
Boonen S, Cheng XG, Nijs J, Nicholson PH, Verbeke G, Lesaffre E, Aerssens J and Dequeker J, 1997. Factors associated with cortical and trabecular bone loss as quantified by peripheral computed tomography (pQCT) at the ultradistal radius in aging women. Calcified Tissue International, 60, 164– 170.
Boonen S, Mohan S, Dequeker J, Aerssens J, Vanderschueren D, Verbeke G, Broos P, Bouillon R and Baylink DJ, 1999. Down-regulation of the serum stimulatory components of the insulin-like growth factor (IGF) system (IGF-I, IGF-II, IGF binding protein [BP]-3, and IGFBP-5) in age-related (type II) femoral neck osteoporosis. Journal of Bone and Mineral Research, 14, 2150– 2158.
Borel P, Caillaud D and Cano NJ, 2015. Vitamin d bioavailability: state of the art. Critical Reviews in Food Science and Nutrition, 55, 1193– 1205.
Bougle D, Sabatier JP, Bureau F, Laroche D, Brouard J, Guillois B and Duhamel JF, 1998. Relationship between bone mineralization and aluminium in the healthy infant. European Journal of Clinical Nutrition, 52, 431– 435.
Bouillon R, Verstuyf A, Mathieu C, Van Cromphaut S, Masuyama R, Dehaes P and Carmeliet G, 2006. Vitamin D resistance. Best Practice and Research. Clinical Endocrinology and Metabolism, 20, 627– 645.
Bouillon R, Carmeliet G, Verlinden L, van Etten E, Verstuyf A, Luderer HF, Lieben L, Mathieu C and Demay M, 2008. Vitamin D and human health: lessons from vitamin D receptor null mice. Endocrine Reviews, 29, 726– 776.
Boxer RS, Dauser DA, Walsh SJ, Hager WD and Kenny AM, 2008. The association between vitamin D and inflammation with the 6-minute walk and frailty in patients with heart failure. Journal of the American Geriatrics Society, 56, 454– 461.
Braam LAJLM, Knapen MHJ, Geusens P, Brouns F, Hamulyak K, Gerichhausen MJW and Vermeer C, 2003. Vitamin K1 supplementation retards bone loss in postmenopausal women between 50 and 60 years of age. Calcified Tissue International, 73, 21– 26.
Braegger C, Campoy C, Colomb V, Decsi T, Domellof M, Fewtrell M, Hojsak I, Mihatsch W, Molgaard C, Shamir R, Turck D and Van Goudoever J on Behalf of the ESPGHAN Committee on Nutrition, 2013. Vitamin D in the healthy European paediatric population. Journal of Pediatric, Gastroenterology, and Nutrition, 56, 692– 701.
Brazier M, Kamel S, Lorget F, Maamer M, Tavera C, Heurtebize N, Grados F, Mathieu M, Garabedian M, Sebert JL and Fardellone P, 2002. Biological effects of supplementation with vitamin D and calcium in postmenopausal women with low bone mass receiving alendronate. Clinical Drug Investigation, 22, 849– 857.
Brembeck P, Winkvist A and Olausson H, 2013. Determinants of vitamin D status in pregnant fair-skinned women in Sweden. British Journal of Nutrition, 110, 856– 864.
Broe KE, Chen TC, Weinberg J, Bischoff-Ferrari HA, Holick MF and Kiel DP, 2007. A higher dose of vitamin d reduces the risk of falls in nursing home residents: a randomized, multiple-dose study. Journal of the American Geriatrics Society, 55, 234– 239.
Brooke OG, Brown IR, Bone CD, Carter ND, Cleeve HJ, Maxwell JD, Robinson VP and Winder SM, 1980. Vitamin D supplements in pregnant Asian women: effects on calcium status and fetal growth. British Medical Journal, 280, 751– 754.
Brouwer-Brolsma EM, Berendsen AAM, Vaes AMM, Dullemeijer C, de Groot LCPGM and Feskens EJM, 2016. Collection and analysis of published scientific information as preparatory work for the setting of dietary reference values for Vitamin D. EFSA supporting publication 2016:EN-766, 171 pp.
Brunner C, Pons-Kühnemann J and Neuhäuser-Berthold M, 2011. Impact of age, anthropometric data and body composition on calcaneal bone characteristics, as measured by quantitative ultrasound (QUS) in an older German population. Ultrasound in Medicine and Biology, 37, 1984– 1992.
Brunvand L and Nordshus T, 1996. Nutritional rickets–an old disease with new relevance. Nordisk Medicin, 111, 219– 221.
Brunvand L, Shah SS, Bergstrom S and Haug E, 1998. Vitamin D deficiency in pregnancy is not associated with obstructed labor. A study among Pakistani women in Karachi. Acta Obstetricia et Gynecologica Scandinavica, 77, 303– 306.
Brustad M, Edvardsen K, Wilsgaard T, Engelsen O, Aksnes L and Lund E, 2007. Seasonality of UV-radiation and vitamin D status at 69 degrees north. Photochemical and Photobiological Sciences, 6, 903– 908.
Bunout D, Barrera G, Leiva L, Gattas V, de la Maza MP, Avendano M and Hirsch S, 2006. Effects of vitamin D supplementation and exercise training on physical performance in Chilean vitamin D deficient elderly subjects. Experimental Gerontology, 41, 746– 752.
Burleigh E, McColl J and Potter J, 2007. Does vitamin D stop inpatients falling? A randomised controlled trial. Age and Ageing, 36, 507– 513.
Burris HH, Rifas-Shiman SL, Camargo CA Jr, Litonjua AA, Huh SY, Rich-Edwards JW and Gillman MW, 2012. Plasma 25-hydroxyvitamin D during pregnancy and small-for-gestational age in black and white infants. Annals of Epidemiology, 22, 581– 586.
Butte NF, Lopez-Alarcon MG and Garza C, 2002. Nutrient Adequacy of Exclusive Breastfeeding for the Term Infant During the First Six Months of Life. World Health Organization, 47 pp.
Cameron ID, Gillespie LD, Robertson MC, Murray GR, Hill KD, Cumming RG and Kerse N, 2012. Interventions for preventing falls in older people in care facilities and hospitals. Cochrane Database of Systematic Reviews, 12, CD005465.
Campbell DE and Fleischman AR, 1988. Rickets of prematurity: controversies in causation and prevention. Clinics in Perinatology, 15, 879– 890.
Cancela L, Le Boulch N and Miravet L, 1986. Relationship between the vitamin D content of maternal milk and the vitamin D status of nursing women and breast-fed infants. Journal of Endocrinology, 110, 43– 50.
Carter GD, 2011. Accuracy of 25-hydroxyvitamin D assays: confronting the issues. Current Drug Targets, 12, 19– 28.
Carter GD, 2012. 25-hydroxyvitamin D: a difficult analyte. Clinical Chemistry, 58, 486– 488.
Cashman KD, 2012. The role of vitamers and dietary-based metabolites of vitamin D in prevention of vitamin D deficiency. Food & Nutrition Research, 2012. 56, 5383 doi:10.3402/fnr.v56i0.5383.
Cashman KD, 2015. Vitamin D: dietary requirements and food fortification as a means of helping achieve adequate vitamin D status. Journal of Steroid Biochemistry and Molecular Biology, 148, 19– 26.
Cashman KD and Kiely M, 2014. Recommended dietary intakes for vitamin D: where do they come from, what do they achieve and how can we meet them? Journal of Human Nutrition and Dietetics, 27, 434– 442.
Cashman KD, Hill TR, Lucey AJ, Taylor N, Seamans KM, Muldowney S, Fitzgerald AP, Flynn A, Barnes MS, Horigan G, Bonham MP, Duffy EM, Strain JJ, Wallace JM and Kiely M, 2008. Estimation of the dietary requirement for vitamin D in healthy adults. American Journal of Clinical Nutrition, 88, 1535– 1542.
Cashman KD, Wallace JM, Horigan G, Hill TR, Barnes MS, Lucey AJ, Bonham MP, Taylor N, Duffy EM, Seamans K, Muldowney S, Fitzgerald AP, Flynn A, Strain JJ and Kiely M, 2009. Estimation of the dietary requirement for vitamin D in free-living adults >=64 y of age. American Journal of Clinical Nutrition, 89, 1366– 1374.
Cashman KD, Fitzgerald AP, Kiely M and Seamans KM, 2011a. A systematic review and meta-regression analysis of the vitamin D intake-serum 25-hydroxyvitamin D relationship to inform European recommendations. British Journal of Nutrition, 106, 1638– 1648.
Cashman KD, FitzGerald AP, Viljakainen HT, Jakobsen J, Michaelsen KF, Lamberg-Allardt C and Molgaard C, 2011b. Estimation of the dietary requirement for vitamin D in healthy adolescent white girls. American Journal of Clinical Nutrition, 93, 549– 555.
Cashman KD, Seamans KM, Lucey AJ, Stocklin E, Weber P, Kiely M and Hill TR, 2012. Relative effectiveness of oral 25-hydroxyvitamin D3 and vitamin D3 in raising wintertime serum 25-hydroxyvitamin D in older adults. American Journal of Clinical Nutrition, 95, 1350– 1356.
Cashman KD, Kiely M, Kinsella M, Durazo-Arvizu RA, Tian L, Zhang Y, Lucey A, Flynn A, Gibney MJ, Vesper HW, Phinney KW, Coates PM, Picciano MF and Sempos CT, 2013. Evaluation of Vitamin D Standardization Program protocols for standardizing serum 25-hydroxyvitamin D data: a case study of the program's potential for national nutrition and health surveys. American Journal of Clinical Nutrition, 97, 1235– 1242.
Cashman KD, Hayes A, O'Donovan SM, Zhang JY, Kinsella M, Galvin K, Kiely M and Seamans KM, 2014. Dietary calcium does not interact with vitamin D3 in terms of determining the response and catabolism of serum 25-hydroxyvitamin D during winter in older adults. American Journal of Clinical Nutrition, 99, 1414– 1423.
Cashman KD, Dowling KG, Skrabakova Z, Gonzalez-Gross M, Valtuena J, De Henauw S, Moreno L, Damsgaard CT, Michaelsen KF, Molgaard C, Jorde R, Grimnes G, Moschonis G, Mavrogianni C, Manios Y, Thamm M, Mensink GB, Rabenberg M, Busch MA, Cox L, Meadows S, Goldberg G, Prentice A, Dekker JM, Nijpels G, Pilz S, Swart KM, van Schoor NM, Lips P, Eiriksdottir G, Gudnason V, Cotch MF, Koskinen S, Lamberg-Allardt C, Durazo-Arvizu RA, Sempos CT and Kiely M, 2016. Vitamin D deficiency in Europe: pandemic? American Journal of Clinical Nutrition, 103, 1033– 1044.
Cauley JA, Lacroix AZ, Wu L, Horwitz M, Danielson ME, Bauer DC, Lee JS, Jackson RD, Robbins JA, Wu C, Stanczyk FZ, LeBoff MS, Wactawski-Wende J, Sarto G, Ockene J and Cummings SR, 2008. Serum 25-hydroxyvitamin D concentrations and risk for hip fractures. Annals of Internal Medicine, 149, 242– 250.
Cauley JA, Parimi N, Ensrud KE, Bauer DC, Cawthon PM, Cummings SR, Hoffman AR, Shikany JM, Barrett-Connor E and Orwoll E and Osteoporotic Fractures in Men Research Group, 2010. Serum 25-hydroxyvitamin D and the risk of hip and nonspine fractures in older men. Journal of Bone and Mineral Research, 25, 545– 553.
Cauley JA, Danielson ME, Boudreau R, Barbour KE, Horwitz MJ, Bauer DC, Ensrud KE, Manson JE, Wactawski-Wende J, Shikany JM and Jackson RD, 2011. Serum 25-hydroxyvitamin D and clinical fracture risk in a multiethnic cohort of women: the Women's Health Initiative (WHI). Journal of Bone and Mineral Research, 26, 2378– 2388.
Cauley JA, Greendale GA, Ruppert K, Lian Y, Randolph JF Jr, Lo JC, Burnett-Bowie SA and Finkelstein JS, 2015. Serum 25 hydroxyvitamin D, bone mineral density and fracture risk across the menopause. Journal of Clinical Endocrinology and Metabolism, 100, 2046– 2054.
Cesur Y, Caksen H, Gundem A, Kirimi E and Odabas D, 2003. Comparison of low and high dose of vitamin D treatment in nutritional vitamin D deficiency rickets. Journal of Pediatric Endocrinology and Metabolism, 16, 1105– 1109.
Chan R, Chan CC, Woo J, Ohlsson C, Mellstrom D, Kwok T and Leung PC, 2011. Serum 25-hydroxyvitamin D, bone mineral density, and non-vertebral fracture risk in community-dwelling older men: results from Mr. Os, Hong Kong. Archives of Osteoporosis, 6, 21– 30.
Chan R, Chan D, Woo J, Ohlsson C, Mellstrom D, Kwok T and Leung PC, 2012. Not all elderly people benefit from vitamin D supplementation with respect to physical function: results from the Osteoporotic Fractures in Men Study, Hong Kong. Journal of the American Geriatrics Society, 60, 290– 295.
Chapuy MC and Meunier PJ, 1997. Vitamin D insufficiency in adults and the elderly. In: D Feldman, FH Glorieux and JW Pike (eds.). Vitamin D. Academic Press, New York, NY, USA. pp. 679– 693.
Chapuy MC, Arlot ME, Duboeuf F, Brun J, Crouzet B, Arnaud S, Delmas PD and Meunier PJ, 1992. Vitamin D3 and calcium to prevent hip fractures in the elderly women. New England Journal of Medicine, 327, 1637– 1642.
Chapuy MC, Pamphile R, Paris E, Kempf C, Schlichting M, Arnaud S, Garnero P and Meunier PJ, 2002. Combined calcium and vitamin D3 supplementation in elderly women: confirmation of reversal of secondary hyperparathyroidism and hip fracture risk: the Decalyos II study. Osteoporosis International, 13, 257– 264.
Cheng S, Lyytikainen A, Kroger H, Lamberg-Allardt C, Alen M, Koistinen A, Wang QJ, Suuriniemi M, Suominen H, Mahonen A, Nicholson PH, Ivaska KK, Korpela R, Ohlsson C, Vaananen KH and Tylavsky F, 2005. Effects of calcium, dairy product, and vitamin D supplementation on bone mass accrual and body composition in 10-12-y-old girls: a 2-y randomized trial. American Journal of Clinical Nutrition, 82, 1115– 1126; quiz 1147-1118.
Chesney RW, Rosen JF, Hamstra AJ, Smith C, Mahaffey K and DeLuca HF, 1981. Absence of seasonal variation in serum concentrations of 1,25-dihydroxyvitamin D despite a rise in 25-hydroxyvitamin D in summer. Journal of Clinical Endocrinology and Metabolism, 53, 139– 142.
Chun RF, Peercy BE, Orwoll ES, Nielson CM, Adams JS and Hewison M, 2014. Vitamin D and DBP: the free hormone hypothesis revisited. Journal of Steroid Biochemistry and Molecular Biology, 144 Pt A, 132– 137.
Chung M, Balk EM, Brendel M, Ip S, Lau J, Lee J, Lichtenstein A, Patel K, Raman G, Tatsioni A, Terasawa T and Trikalinos TA, 2009. Vitamin D and calcium: a systematic review of health outcomes. Evidence Report/Technology Assessment, 183, 1– 420.
Clausen I, Jakobsen J, Leth T and Ovesen L, 2003. Vitamin D3 and 25-hydroxyvitamin D3 in raw and cooked pork cuts. Journal of Food Composition and Analysis, 16, 575– 585.
Close GL, Russell J, Cobley JN, Owens DJ, Wilson G, Gregson W, Fraser WD and Morton JP, 2013a. Assessment of vitamin D concentration in non-supplemented professional athletes and healthy adults during the winter months in the UK: implications for skeletal muscle function. Journal of Sports Sciences, 31, 344– 353.
Close GL, Leckey J, Patterson M, Bradley W, Owens DJ, Fraser WD and Morton JP, 2013b. The effects of vitamin D3 supplementation on serum total 25[OH]D concentration and physical performance: a randomised dose–response study. British Journal of Sports Medicine, 47, 692– 696.
Cockburn F, Belton NR, Purvis RJ, Giles MM, Brown JK, Turner TL, Wilkinson EM, Forfar JO, Barrie WJ, McKay GS and Pocock SJ, 1980. Maternal vitamin D intake and mineral metabolism in mothers and their newborn infants. British Medical Journal, 281, 11– 14.
Congdon P, Horsman A, Kirby PA, Dibble J and Bashir T, 1983. Mineral content of the forearms of babies born to Asian and white mothers. British Medical Journal (Clinical Research Edition), 286, 1233– 1235.
Cooper C, McLaren M, Wood PJ, Coulton L and Kanis JA, 1989. Indices of calcium metabolism in women with hip fractures. Bone and Mineral, 5, 193– 200.
Cooper L, Clifton-Bligh PB, Nery ML, Figtree G, Twigg S, Hibbert E and Robinson BG, 2003. Vitamin D supplementation and bone mineral density in early postmenopausal women. American Journal of Clinical Nutrition, 77, 1324– 1329.
Cooper C, Harvey NC, Bishop NJ, Kennedy S, Papageorghiou AT, Schoenmakers I, Fraser R, Gandhi SV, Carr A, D'Angelo S, Crozier SR, Moon RJ, Arden NK, Dennison EM, Godfrey KM, Inskip HM, Prentice A, Mughal MZ, Eastell R, Reid DM and Javaid MK and Group MS, 2016. Maternal gestational vitamin D supplementation and offspring bone health (MAVIDOS): a multicentre, double-blind, randomised placebo-controlled trial. Lancet Diabetes Endocrinology, 4, 393– 402.
Cranney A, Horsley T, O'Donnell S, Weiler H, Puil L, Ooi D, Atkinson S, Ward L, Moher D, Hanley D, Fang M, Yazdi F, Garritty C, Sampson M, Barrowman N, Tsertsvadze A and Mamaladze V, 2007. Effectiveness and safety of vitamin D in relation to bone health. Evidence Report/Technology Assessment, 158, 1– 235.
Cross NA, Hillman LS, Allen SH, Krause GF and Vieira NE, 1995. Calcium homeostasis and bone metabolism during pregnancy, lactation, and postweaning: a longitudinal study. American Journal of Clinical Nutrition, 61, 514– 523.
Cummings SR, Browner WS, Bauer D, Stone K, Ensrud K, Jamal S and Ettinger B, 1998. Endogenous hormones and the risk of hip and vertebral fractures among older women. Study of Osteoporotic Fractures Research Group. New England Journal of Medicine, 339, 733– 738.
Cynober L, Alix E, Arnaud-Battandier F, Bonnefoy M, Brocker P, Cals MJ, Cherbut C, Coplo C, Ferry M, Ghisolfi-Marque A, Kravtchenko T, Lesourd B, Mignot C and Patureau Mirand P, 2000. Apports nutritionnels conseillés chez la personne âgée. Nutrition Clinique et Métabolisme, 14, 1S– 14S.
D-A-CH (Deutsche Gesellschaft für Ernährung, Österreichische Gesellschaft für Ernährung, Schweizerische Gesellschaft für Ernährung), 2015. Referenzwerte für die Nährstoffzufuhr. 2. Auflage/1. Ausgabe, DGE, Bonn, Germany.
Dam TT, von Muhlen D and Barrett-Connor EL, 2009. Sex-specific association of serum vitamin D levels with physical function in older adults. Osteoporosis International, 20, 751– 760.
Darling AL, Hart KH, MacDonald HM, Horton K, Kang'Ombe AR, Berry JL and Lanham-New SA, 2013. Vitamin D deficiency in UK South Asian Women of childbearing age: a comparative longitudinal investigation with UK Caucasian women. Osteoporosis International, 24, 477– 488.
Datta S, Alfaham M, Davies DP, Dunstan F, Woodhead S, Evans J and Richards B, 2002. Vitamin D deficiency in pregnant women from a non-European ethnic minority population–an interventional study. British Journal of Obstetrics and Gynaecology, 109, 905– 908.
Dawodu A and Tsang RC, 2012. Maternal vitamin D status: effect on milk vitamin D content and vitamin D status of breastfeeding infants. Advances in Nutrition, 3, 353– 361.
Dawodu A, Agarwal M, Sankarankutty M, Hardy D, Kochiyil J and Badrinath P, 2005. Higher prevalence of vitamin D deficiency in mothers of rachitic than nonrachitic children. Journal of Pediatrics, 147, 109– 111.
Dawson-Hughes B, 1996. Calcium and vitamin D nutritional needs of elderly women. Journal of Nutrition, 126, 1165S– 1167S.
Dawson-Hughes B, Dallal GE, Krall EA, Harris S, Sokoll LJ and Falconer G, 1991. Effect of vitamin D supplementation on wintertime and overall bone loss in healthy postmenopausal women. Annals of Internal Medicine, 115, 505– 512.
Dawson-Hughes B, Harris SS, Krall EA, Dallal GE, Falconer G and Green CL, 1995. Rates of bone loss in postmenopausal women randomly assigned to one of two dosages of vitamin D. American Journal of Clinical Nutrition, 61, 1140– 1145.
Dawson-Hughes B, Harris SS, Krall EA and Dallal GE, 1997. Effect of calcium and vitamin D supplementation on bone density in men and women 65 years of age or older. New England Journal of Medicine, 337, 670– 676.
Dawson-Hughes B, Heaney RP, Holick MF, Lips P, Meunier PJ and Vieth R, 2005. Estimates of optimal vitamin D status. Osteoporosis International, 16, 713– 716.
Dawson-Hughes B, Mithal A, Bonjour JP, Boonen S, Burckhardt P, Fuleihan GE, Josse RG, Lips P, Morales-Torres J and Yoshimura N, 2010. IOF position statement: vitamin D recommendations for older adults. Osteoporosis International, 21, 1151– 1154.
DeLappe E, McGreevy C, ni Chadhain N, Grimes H, O'Brien T and Mulkerrin E, 2006. Vitamin D insufficiency in older female community-dwelling acute hospital admissions and the response to supplementation. European Journal of Clinical Nutrition, 60, 1009– 1015.
Delvin EE, Salle BL, Glorieux FH, Adeleine P and David LS, 1986. Vitamin D supplementation during pregnancy: effect on neonatal calcium homeostasis. Journal of Pediatrics, 109, 328– 334.
Dennison E, Eastell R, Fall CH, Kellingray S, Wood PJ and Cooper C, 1999. Determinants of bone loss in elderly men and women: a prospective population-based study. Osteoporosis International, 10, 384– 391.
De-Regil LM, Palacios C, Ansary A, Kulier R and Pena-Rosas JP, 2012. Vitamin D supplementation for women during pregnancy. Cochrane Database of Systematic Reviews, 2, CD008873.
DerSimonian R and Laird N, 1986. Meta-analysis in clinical trials. Controlled Clinical Trials, 7, 177– 188.
Devine A, Wilson SG, Dick IM and Prince RL, 2002. Effects of vitamin D metabolites on intestinal calcium absorption and bone turnover in elderly women. American Journal of Clinical Nutrition, 75, 283– 288.
DH (Department of Health), 1991. Dietary reference values for food energy and nutrients for the United Kingdom. Report of the Panel on dietary reference values of the Committee on Medical Aspects of Food Policy. HM Stationary Office, London, UK, 212 pp.
DH (Department of Health), 1998. Nutrition and bone health: with particular reference to calcium and vitamin D. Great Britain Committee on Medical Aspects of Food Policy, Working Group on the Nutritional Status of the Population. Report on Health and Social Subjects (49). TSO (The Stationery Office), London, UK, 124 pp.
Dhesi JK, Jackson SH, Bearne LM, Moniz C, Hurley MV, Swift CG and Allain TJ, 2004. Vitamin D supplementation improves neuromuscular function in older people who fall. Age and Ageing, 33, 589– 595.
Diamond T, Smerdely P, Kormas N, Sekel R, Vu T and Day P, 1998. Hip fracture in elderly men: the importance of subclinical vitamin D deficiency and hypogonadism. Medical Journal of Australia, 169, 138– 141.
Dinour D, Beckerman P, Ganon L, Tordjman K, Eisenstein Z and Holtzman EJ, 2013. Loss-of-function mutations of CYP24A1, the vitamin D 24-hydroxylase gene, cause long-standing hypercalciuric nephrolithiasis and nephrocalcinosis. Journal of Urology, 190, 552– 557.
Drincic AT, Armas LA, Van Diest EE and Heaney RP, 2012. Volumetric dilution, rather than sequestration best explains the low vitamin D status of obesity. Obesity (Silver Spring), 20, 1444– 1448.
Dror DK, King JC, Fung EB, Van Loan MD, Gertz ER and Allen LH, 2012. Evidence of associations between feto-maternal vitamin D status, cord parathyroid hormone and bone-specific alkaline phosphatase, and newborn whole body bone mineral content. Nutrients, 4, 68– 77.
Edelmann-Schafer B, Berthold LD, Stracke H, Luhrmann PM and Neuhäuser-Berthold M, 2011. Identifying elderly women with osteoporosis by spinal dual X-ray absorptiometry, calcaneal quantitative ultrasound and spinal quantitative computed tomography: a comparative study. Ultrasound in Medicine and Biology, 37, 29– 36.
EFSA (European Food Safety Authority), 2016. Outcome of a public consultation on the Draft Scientific Opinion of the EFSA Panel on Dietetic Products, Nutrition and Allergies (NDA) on dietary reference values for vitamin D. EFSA supporting publication 2016:EN-1078, 100 pp.
EFSA NDA Panel (EFSA Panel on Dietetic Products, Nutrition and Allergies), 2009. Scientific Opinion on the appropriate age for introduction of complementary feeding of infants. EFSA Journal 2009; 7(12): 1423, 38 pp. doi:10.2903/j.efsa.2009.1423
EFSA NDA Panel (EFSA Panel on Dietetic Products, Nutrition and Allergies), 2010. Scientific Opinion on principles for deriving and applying dietary reference values. EFSA Journal 2010; 8(3): 1458, 30 pp. doi:10.2903/j.efsa.2010.1458
EFSA NDA Panel (EFSA Panel on Dietetic Products, Nutrition and Allergies), 2011. Scientific Opinion on the substantiation of a health claim related to vitamin D and risk of falling pursuant to Article 14 of Regulation (EC) No1924/2006. EFSA Journal 2011; 9(9): 2382, 18 pp. doi:10.2903/j.efsa.2011.2382
EFSA NDA Panel (EFSA Panel on Dietetic Products, Nutrition and Allergies), 2012a. Scientific Opinion on the Tolerable Upper Intake Level of vitamin D. EFSA Journal 2012; 10(7): 2813, 45 pp. doi:10.2903/j.efsa.2012.2813
EFSA NDA Panel (EFSA Panel on Dietetic Products, Nutrition and Allergies), 2012b. Guidance on the scientific requirements for health claims related to bone, joints, skin, and oral health. EFSA Journal 2012; 10(5): 2702, 13 pp. doi:10.2903/j.efsa.2012.2702
EFSA NDA Panel (EFSA Panel on Dietetic Products Nutrition and Allergies), 2013. Scientific Opinion on nutrient requirements and dietary intakes of infants and young children in the European Union. EFSA Journal 2013; 11(10): 3408, 103 pp. doi:10.2903/j.efsa.2013.3408
EFSA NDA Panel (EFSA Panel on Dietetic Products, Nutrition and Allergies), 2015a. Scientific Opinion on dietary reference values for vitamin A. EFSA Journal 2015; 13(3): 4028, 84 pp. doi:10.2903/j.efsa.2015.4028
EFSA NDA Panel (EFSA Panel on Dietetic Products, Nutrition and Allergies), 2015b. Scientific Opinion on dietary reference values for calcium. EFSA Journal 2015; 13(5): 4101, 82 pp. doi:10.2903/j.efsa.2015.4101
EFSA NDA Panel (EFSA Panel on Dietetic Products, Nutrition and Allergies), 2015c. Scientific Opinion on dietary reference values for phosphorus. EFSA Journal 2015; 13(7): 4185, 54 pp. doi:10.2903/j.efsa.2015.4185
Egger M, Davey Smith G, Schneider M and Minder C, 1997. Bias in meta-analysis detected by a simple, graphical test. British Medical Journal, 315, 629– 634.
El-Hajj Fuleihan G, Nabulsi M, Tamim H, Maalouf J, Salamoun M, Khalife H, Choucair M, Arabi A and Vieth R, 2006. Effect of vitamin D replacement on musculoskeletal parameters in school children: a randomized controlled trial. Journal of Clinical Endocrinology and Metabolism, 91, 405– 412.
Elzouki AY, Markestad T, Elgarrah M, Elhoni N and Aksnes L, 1989. Serum concentrations of vitamin D metabolites in rachitic Libyan children. Journal of Pediatric Gastroenterology and Nutrition, 9, 507– 512.
Engelman CD, Meyers KJ, Iyengar SK, Liu Z, Karki CK, Igo RP Jr, Truitt B, Robinson J, Sarto GE, Wallace R, Blodi BA, Klein ML, Tinker L, LeBlanc ES, Jackson RD, Song Y, Manson JE, Mares JA and Millen AE, 2013. Vitamin D intake and season modify the effects of the GC and CYP2R1 genes on 25-hydroxyvitamin D concentrations. Journal of Nutrition, 143, 17– 26.
Engelsen O, 2010. The relationship between ultraviolet radiation exposure and vitamin D status. Nutrients, 2, 482– 495.
Engelsen O, Brustad M, Aksnes L and Lund E, 2005. Daily duration of vitamin D synthesis in human skin with relation to latitude, total ozone, altitude, ground cover, aerosols and cloud thickness. Photochemistry and Photobiology, 81, 1287– 1290.
Ensrud KE, Taylor BC, Paudel ML, Cauley JA, Cawthon PM, Cummings SR, Fink HA, Barrett-Connor E, Zmuda JM, Shikany JM and Orwoll ES and Osteoporotic Fractures in Men Study Group, 2009. Serum 25-hydroxyvitamin D levels and rate of hip bone loss in older men. Journal of Clinical Endocrinology and Metabolism, 94, 2773– 2780.
Erem C, Tanakol R, Alagol F, Omer B and Cetin O, 2002. Relationship of bone turnover parameters, endogenous hormones and vit D deficiency to hip fracture in elderly postmenopausal women. International Journal of Clinical Practice, 56, 333– 337.
ESVITAF, 1986. Vitamin status in three groups of French adults: controls, obese subjects, alcohol drinkers. Annals of Nutrition and Metabolism, 30(Suppl 1), 1– 94.
FAO/WHO/UNU (Food and Agriculture Organization of the United Nations/World Health Organization/United Nations University), 2004. Human energy requirements. Report of a Joint FAO/WHO/UNU Expert Consultation: Rome, 17–24 October 2001. FAO Food and Nutrition Technical Report Series, 103 pp.
Farrant HJ, Krishnaveni GV, Hill JC, Boucher BJ, Fisher DJ, Noonan K, Osmond C, Veena SR and Fall CH, 2009. Vitamin D insufficiency is common in Indian mothers but is not associated with gestational diabetes or variation in newborn size. European Journal of Clinical Nutrition, 63, 646– 652.
Farrell CJ, Martin S, McWhinney B, Straub I, Williams P and Herrmann M, 2012. State-of-the-art vitamin D assays: a comparison of automated immunoassays with liquid chromatography-tandem mass spectrometry methods. Clinical Chemistry, 58, 531– 542.
Fernandez-Alonso AM, Dionis-Sanchez EC, Chedraui P, Gonzalez-Salmeron MD and Perez-Lopez FR, Spanish Vitamin D and Women's Health Research Group, 2012. First-trimester maternal serum 25-hydroxyvitamin D(3) status and pregnancy outcome. International Journal of Gynaecology and Obstetrics, 116, 6– 9.
Flicker L, MacInnis RJ, Stein MS, Scherer SC, Mead KE, Nowson CA, Thomas J, Lowndes C, Hopper JL and Wark JD, 2005. Should older people in residential care receive vitamin D to prevent falls? Results of a randomized trial. Journal of the American Geriatrics Society, 53, 1881– 1888.
Fomon SJ, Younoszai MK and Thomas LN, 1966. Influence of vitamin D on linear growth of normal full-term infants. Journal of Nutrition, 88, 345– 350.
Forman JP, Scott JB, Ng K, Drake BF, Suarez E, Hayden DL, Bennett GG, Chandler PD, Hollis BW, Emmons KM, Giovannucci EL, Fuchs CS and Chan AT, 2013. Effect of vitamin D supplementation on blood pressure in blacks. Hypertension, 61, 779– 785.
Francis RM, Boyle IT, Moniz C, Sutcliffe AM, Davis BS, Beastall GH, Cowan RA and Downes N, 1996. A comparison of the effects of alfacalcidol treatment and vitamin D2 supplementation on calcium absorption in elderly women with vertebral fractures. Osteoporosis International, 6, 284– 290.
Frolich A, Rudnicki M, Storm T, Rasmussen N and Hegedus L, 1992. Impaired 1,25-dihydroxyvitamin D production in pregnancy-induced hypertension. European Journal of Obstetrics, Gynecology, and Reproductive Biology, 47, 25– 29.
Gale CR, Robinson SM, Harvey NC, Javaid MK, Jiang B, Martyn CN, Godfrey KM and Cooper C and Princess Anne Hospital Study Group, 2008. Maternal vitamin D status during pregnancy and child outcomes. European Journal of Clinical Nutrition, 62, 68– 77.
Gallagher JC, Yalamanchili V and Smith LM, 2012. The effect of vitamin D on calcium absorption in older women. Journal of Clinical Endocrinology and Metabolism, 97, 3550– 3556.
Gallagher JC, Jindal PS and Smith LM, 2014. Vitamin D does not increase calcium absorption in young women: a randomized clinical trial. Journal of Bone and Mineral Research, 29, 1081– 1087.
Gallo S, Comeau K, Vanstone C, Agellon S, Sharma A, Jones G, L'Abbe M, Khamessan A, Rodd C and Weiler H, 2013. Effect of different dosages of oral vitamin D supplementation on vitamin D status in healthy, breastfed infants: a randomized trial. JAMA, 309, 1785– 1792.
Garabedian M, Vainsel M, Mallet E, Guillozo H, Toppet M, Grimberg R, Nguyen TM and Balsan S, 1983. Circulating vitamin-D metabolite concentrations in children with nutritional rickets. Journal of Pediatrics, 103, 381– 386.
Garabedian M, Zeghoud F and Rossignol C, 1991. Les besoins en vitamine D du nourrisson vivant en France. In: Journées Parisiennes de Pédiatrie. Médecine-Sciences, Flammarion, Paris, France. pp. 51– 57.
Gaugris S, Heaney RP, Boonen S, Kurth H, Bentkover JD and Sen SS, 2005. Vitamin D inadequacy among post-menopausal women: a systematic review. QJM: Monthly Journal of the Association of Physicians, 98, 667– 676.
Gerdhem P, Ringsberg KA, Obrant KJ and Akesson K, 2005. Association between 25-hydroxy vitamin D levels, physical activity, muscle strength and fractures in the prospective population-based OPRA Study of Elderly Women. Osteoporosis International, 16, 1425– 1431.
Gernand AD, Simhan HN, Klebanoff MA and Bodnar LM, 2013. Maternal serum 25-hydroxyvitamin D and measures of newborn and placental weight in a U.S. multicenter cohort study. Journal of Clinical Endocrinology and Metabolism, 98, 398– 404.
Gessner BD, deSchweinitz E, Petersen KM and Lewandowski C, 1997. Nutritional rickets among breast-fed black and Alaska Native children. Alaska Medicine, 39, 72– 74, 87.
Ghannam NN, Hammami MM, Bakheet SM and Khan BA, 1999. Bone mineral density of the spine and femur in healthy Saudi females: relation to vitamin D status, pregnancy, and lactation. Calcified Tissue International, 65, 23– 28.
Ghazi AA, Hosseinpanah F, M Ardakani E, Ghazi S, Hedayati M and Azizi F, 2010. Effects of different doses of oral cholecalciferol on serum 25(OH)D, PTH, calcium and bone markers during fall and winter in schoolchildren. European Journal of Clinical Nutrition, 64, 1415– 1422.
Gifre L, Peris P, Monegal A, Martinez de Osaba MJ, Alvarez L and Guanabens N, 2011. Osteomalacia revisited: a report on 28 cases. Clinical Rheumatology, 30, 639– 645.
Gillespie LD, Robertson MC, Gillespie WJ, Sherrington C, Gates S, Clemson LM and Lamb SE, 2012. Interventions for preventing falls in older people living in the community. Cochrane Database of Systematic Reviews, 9, CD007146.
Glerup H, Mikkelsen K, Poulsen L, Hass E, Overbeck S, Thomsen J, Charles P and Eriksen EF, 2000. Commonly recommended daily intake of vitamin D is not sufficient if sunlight exposure is limited. Journal of Internal Medicine, 247, 260– 268.
Goussous R, Song L, Dallal GE and Dawson-Hughes B, 2005. Lack of effect of calcium intake on the 25-hydroxyvitamin D response to oral vitamin D3. Journal of Clinical Endocrinology and Metabolism, 90, 707– 711.
Graafmans WC, Ooms ME, Hofstee HM, Bezemer PD, Bouter LM and Lips P, 1996. Falls in the elderly: a prospective study of risk factors and risk profiles. American Journal of Epidemiology, 143, 1129– 1136.
Grados F, Brazier M, Kamel S, Duver S, Heurtebize N, Maamer M, Mathieu M, Garabedian M, Sebert JL and Fardellone P, 2003. Effects on bone mineral density of calcium and vitamin D supplementation in elderly women with vitamin D deficiency. Joint, Bone, Spine, 70, 203– 208.
Graff M, Thacher TD, Fischer PR, Stadler D, Pam SD, Pettifor JM, Isichei CO and Abrams SA, 2004. Calcium absorption in Nigerian children with rickets. American Journal of Clinical Nutrition, 80, 1415– 1421.
Grant AM, Avenell A, Campbell MK, McDonald AM, MacLennan GS, McPherson GC, Anderson FH, Cooper C, Francis RM, Donaldson C, Gillespie WJ, Robinson CM, Torgerson DJ and Wallace WA and Group RT, 2005. Oral vitamin D3 and calcium for secondary prevention of low-trauma fractures in elderly people (Randomised Evaluation of Calcium Or vitamin D, RECORD): a randomised placebo-controlled trial. Lancet, 365, 1621– 1628.
Greer FR and Marshall S, 1989. Bone mineral content, serum vitamin D metabolite concentrations, and ultraviolet B light exposure in infants fed human milk with and without vitamin D2 supplements. Journal of Pediatrics, 114, 204– 212.
Greer FR, Searcy JE, Levin RS, Steichen JJ, Asch PS and Tsang RC, 1981. Bone mineral content and serum 25-hydroxyvitamin D concentration in breast-fed infants with and without supplemental vitamin D. Journal of Pediatrics, 98, 696– 701.
Greer FR, Searcy JE, Levin RS, Steichen JJ, Steichen-Asche PS and Tsang RC, 1982. Bone mineral content and serum 25-hydroxyvitamin D concentrations in breast-fed infants with and without supplemental vitamin D: one-year follow-up. Journal of Pediatrics, 100, 919– 922.
Gropper S, Smith J and Groff J, 2009. Advanced Nutrition & Human Metabolism. Wadsworth Cengage Learning, Belmont, MA, USA. 587 pp.
Grossmann RE and Tangpricha V, 2010. Evaluation of vehicle substances on vitamin D bioavailability: a systematic review. Molecular Nutrition and Food Research, 54, 1055– 1061.
de Gruijl FR and Pavel S, 2012. The effects of a mid-winter 8-week course of sub-sunburn sunbed exposures on tanning, vitamin D status and colds. Photochemical and Photobiological Sciences, 11, 1848– 1854.
Gultekin A, Ozalp I, Hasanoglu A and Unal A, 1987. Serum-25-hydroxycholecalciferol levels in children and adolescents. Turkish Journal of Pediatrics, 29, 155– 162.
Haliloglu B, Ilter E, Aksungar FB, Celik A, Coksuer H, Gunduz T, Yucel E and Ozekici U, 2011. Bone turnover and maternal 25(OH) vitamin D3 levels during pregnancy and the postpartum period: should routine vitamin D supplementation be increased in pregnant women? European Journal of Obstetrics, Gynecology, and Reproductive Biology, 158, 24– 27.
Hansen KE, Jones AN, Lindstrom MJ, Davis LA, Engelke JA and Shafer MM, 2008. Vitamin D insufficiency: disease or no disease? Journal of Bone and Mineral Research, 23, 1052– 1060.
Hansen AL, Dahl L, Bakke L, Froyland L and Thayer JF, 2010. Fish consumption and heart rate variability: preliminary results. Journal of Psychophysiology, 24, 41– 47.
Harris SS and Dawson-Hughes B, 2002. Plasma vitamin D and 25OHD responses of young and old men to supplementation with vitamin D3. Journal of the American College of Nutrition, 21, 357– 362.
Harvey NC, Holroyd C, Ntani G, Javaid K, Cooper P, Moon R, Cole Z, Tinati T, Godfrey K, Dennison E, Bishop NJ, Baird J and Cooper C, 2014. Vitamin D supplementation in pregnancy: a systematic review. Health Technology Assessment, 18, 1– 190.
Harwood RH, Sahota O, Gaynor K, Masud T and Hosking DJ and Nottingham Neck of Femur Study, 2004. A randomised, controlled comparison of different calcium and vitamin D supplementation regimens in elderly women after hip fracture: the Nottingham Neck of Femur (NONOF) Study. Age and Ageing, 33, 45– 51.
Hathcock JN, Shao A, Vieth R and Heaney R, 2007. Risk assessment for vitamin D. American Journal of Clinical Nutrition, 85, 6– 18.
Haugen M, Brantsaeter AL, Trogstad L, Alexander J, Roth C, Magnus P and Meltzer HM, 2009. Vitamin D supplementation and reduced risk of preeclampsia in nulliparous women. Epidemiology, 20, 720– 726.
Hazell TJ, Pham TT, Jean-Philippe S, Finch SL, El Hayek J, Vanstone CA, Agellon S, Rodd CJ and Weiler HA, 2015. Vitamin D status is associated with bone mineral density and bone mineral content in preschool-aged children. Journal of Clinical Densitometry, 18, 60– 67.
Health Council of the Netherlands, 2000. Dietary reference intakes: calcium, vitamin D, thiamin, riboflavin, niacin, pantothenic acid, and biotin. Publication no. 2000/12. 180 pp.
Health Council of the Netherlands, 2012. Evaluation of dietary reference values for vitamin D. No 2012/15E, 138 pp.
Heaney RP, 2000. Dietary protein and phosphorus do not affect calcium absorption. American Journal of Clinical Nutrition, 72, 758– 761.
Heaney RP, Abrams S, Dawson-Hughes B, Looker A, Marcus R, Matkovic V and Weaver C, 2000. Peak bone mass. Osteoporosis International, 11, 985– 1009.
Heaney RP, Dowell MS, Hale CA and Bendich A, 2003a. Calcium absorption varies within the reference range for serum 25-hydroxyvitamin D. Journal of the American College of Nutrition, 22, 142– 146.
Heaney RP, Davies KM, Chen TC, Holick MF and Barger-Lux MJ, 2003b. Human serum 25-hydroxycholecalciferol response to extended oral dosing with cholecalciferol. American Journal of Clinical Nutrition, 77, 204– 210.
Heaney RP, Armas LA, Shary JR, Bell NH, Binkley N and Hollis BW, 2008. 25-Hydroxylation of vitamin D3: relation to circulating vitamin D3 under various input conditions. American Journal of Clinical Nutrition, 87, 1738– 1742.
Heaney RP, Horst RL, Cullen DM and Armas LA, 2009. Vitamin D3 distribution and status in the body. Journal of the American College of Nutrition, 28, 252– 256.
Heijboer AC, Blankenstein MA, Kema IP and Buijs MM, 2012. Accuracy of 6 routine 25-hydroxyvitamin D assays: influence of vitamin D binding protein concentration. Clinical Chemistry, 58, 543– 548.
Heikkinen A, Parviainen MT, Tuppurainen MT, Niskanen L, Komulainen MH and Saarikoski S, 1998. Effects of postmenopausal hormone replacement therapy with and without vitamin D3 on circulating levels of 25-hydroxyvitamin D and 1,25-dihydroxyvitamin D. Calcified Tissue International, 62, 26– 30.
Hercberg S, Preziosi P, Galan P, Devanlay M, Keller H, Bourgeois C, Potier de Courcy G and Cherouvrier F, 1994. Vitamin status of a healthy French population: dietary intakes and biochemical markers. International Journal for Vitamin and Nutrition Research, 64, 220– 232.
Higgins JPT and Green S, 2011. Cochrane Handbook for Systematic Reviews of Interventions. Version 5.1.0 [updated March 2011]. The Cochrane Collaboration. Available from www.cochrane-handbook.org
Higgins JP and Thompson SG, 2002. Quantifying heterogeneity in a meta-analysis. Statistics in Medicine, 21, 1539– 1558.
Higgins JP and Thompson SG, 2004. Controlling the risk of spurious findings from meta-regression. Statistics in Medicine, 23, 1663– 1682.
Hill T, Collins A, O'Brien M, Kiely M, Flynn A and Cashman KD, 2005. Vitamin D intake and status in Irish postmenopausal women. European Journal of Clinical Nutrition, 59, 404– 410.
Holick MF, 1994. McCollum Award Lecture, 1994: vitamin D–new horizons for the 21st century. American Journal of Clinical Nutrition, 60, 619– 630.
Holick MF, 1998. Vitamin D requirements for humans of all ages: new increased requirements for women and men 50 years and older. Osteoporosis International, 8(Suppl 2), S24– S29.
Holick MF, 1999. Vitamin D: Physiology, Molecular Biology, and Clinical Applications. Humana Press, Totowa, NJ, USA. 458 pp.
Holick MF, 2004. Vitamin D: importance in the prevention of cancers, type 1 diabetes, heart disease, and osteoporosis. American Journal of Clinical Nutrition, 79, 362– 371.
Holick MF, 2006. Resurrection of vitamin D deficiency and rickets. Journal of Clinical Investigation, 116, 2062– 2072.
Holick MF, 2007. Vitamin D deficiency. New England Journal of Medicine, 357, 266– 281.
Holick MF, Biancuzzo RM, Chen TC, Klein EK, Young A, Bibuld D, Reitz R, Salameh W, Ameri A and Tannenbaum AD, 2008. Vitamin D2 is as effective as vitamin D3 in maintaining circulating concentrations of 25-hydroxyvitamin D. Journal of Clinical Endocrinology and Metabolism, 93, 677– 681.
Holick MF, Binkley NC, Bischoff-Ferrari HA, Gordon CM, Hanley DA, Heaney RP, Murad MH and Weaver CM, 2012. Guidelines for preventing and treating vitamin D deficiency and insufficiency revisited. Journal of Clinical Endocrinology and Metabolism, 97, 1153– 1158.
Hollis BW and Wagner CL, 2004a. Vitamin D requirements during lactation: high-dose maternal supplementation as therapy to prevent hypovitaminosis D for both the mother and the nursing infant. American Journal of Clinical Nutrition, 80, 1752S– 1758S.
Hollis BW and Wagner CL, 2004b. Assessment of dietary vitamin D requirements during pregnancy and lactation. American Journal of Clinical Nutrition, 79, 717– 726.
Hollis BW, Pittard WB 3rd and Reinhardt TA, 1986. Relationships among vitamin D, 25-hydroxyvitamin D, and vitamin D-binding protein concentrations in the plasma and milk of human subjects. Journal of Clinical Endocrinology and Metabolism, 62, 41– 44.
Hollis BW, Johnson D, Hulsey TC, Ebeling M and Wagner CL, 2011. Vitamin D supplementation during pregnancy: double-blind, randomized clinical trial of safety and effectiveness. Journal of Bone and Mineral Research, 26, 2341– 2357.
Holm L, Olesen JL, Matsumoto K, Doi T, Mizuno M, Alsted TJ, Mackey AL, Schwarz P and Kjaer M, 2008. Protein-containing nutrient supplementation following strength training enhances the effect on muscle mass, strength, and bone formation in postmenopausal women. Journal of Applied Physiology, 105, 274– 281.
Holmberg I, Berlin T, Ewerth S and Bjorkhem I, 1986. 25-Hydroxylase activity in subcellular fractions from human liver. Evidence for different rates of mitochondrial hydroxylation of vitamin D2 and D3. Scandinavian Journal of Clinical and Laboratory Investigation, 46, 785– 790.
Holmlund-Suila E, Viljakainen H, Hytinantti T, Lamberg-Allardt C, Andersson S and Makitie O, 2012. High-dose vitamin d intervention in infants–effects on vitamin d status, calcium homeostasis, and bone strength. Journal of Clinical Endocrinology and Metabolism, 97, 4139– 4147.
Holt E and Wysolmerski JJ, 2011. Parathyroid hormone, parathyroid hormone-related protein, and calcitonin. In: D Feldman, JW Pike and JS Adams (eds.). Vitamin D. Academic Press, San Diego, CA, USA. pp. 725– 745.
Holvik K, Ahmed LA, Forsmo S, Gjesdal CG, Grimnes G, Samuelsen SO, Schei B, Blomhoff R, Tell GS and Meyer HE, 2013. Low serum levels of 25-hydroxyvitamin D predict hip fracture in the elderly: a NOREPOS study. Journal of Clinical Endocrinology and Metabolism, 98, 3341– 3350.
Honkanen R, Alhava E, Parviainen M, Talasniemi S and Monkkonen R, 1990. The necessity and safety of calcium and vitamin D in the elderly. Journal of the American Geriatrics Society, 38, 862– 866.
Houston DK, Tooze JA, Davis CC, Chaves PH, Hirsch CH, Robbins JA, Arnold AM, Newman AB and Kritchevsky SB, 2011. Serum 25-hydroxyvitamin D and physical function in older adults: the Cardiovascular Health Study All Stars. Journal of the American Geriatrics Society, 59, 1793– 1801.
Houston DK, Tooze JA, Neiberg RH, Hausman DB, Johnson MA, Cauley JA, Bauer DC, Cawthon PM, Shea MK, Schwartz GG, Williamson JD, Tylavsky FA, Visser M, Simonsick EM, Harris TB, Kritchevsky SB and Health ABCS, 2012. 25-hydroxyvitamin D status and change in physical performance and strength in older adults: the Health, Aging, and Body Composition Study. American Journal of Epidemiology, 176, 1025– 1034.
Hower J, Knoll A, Ritzenthaler KL, Steiner C and Berwind R, 2013. Vitamin D fortification of growing up milk prevents decrease of serum 25-hydroxyvitamin D concentrations during winter: a clinical intervention study in Germany. European Journal of Pediatrics, 172, 1597– 1605.
Hunter D, Major P, Arden N, Swaminathan R, Andrew T, MacGregor AJ, Keen R, Snieder H and Spector TD, 2000. A randomized controlled trial of vitamin D supplementation on preventing postmenopausal bone loss and modifying bone metabolism using identical twin pairs. Journal of Bone and Mineral Research, 15, 2276– 2283.
Hypponen E, Hartikainen AL, Sovio U, Jarvelin MR and Pouta A, 2007. Does vitamin D supplementation in infancy reduce the risk of pre-eclampsia? European Journal of Clinical Nutrition, 61, 1136– 1139.
Hypponen E, Berry DJ, Wjst M and Power C, 2009. Serum 25-hydroxyvitamin D and IgE – a significant but nonlinear relationship. Allergy, 64, 613– 620.
Hyppönen E, Fararouei M, Sovio U, Hartikainen AL, Pouta A, Robertson C, Whittaker JC and Jarvelin MR, 2011. High-dose vitamin D supplements are not associated with linear growth in a large Finnish cohort. Journal of Nutrition, 141, 843– 848.
IOM (Institute of Medicine), 1997. Dietary Reference Intakes for Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride. National Academy Press, Washington, DC, USA. 454 pp.
IOM (Institute of Medicine), 2011. Dietary Reference Intakes for Calcium and Vitamin D. National Academies Press, Washington, DC, USA. 1115 pp.
Islam MZ, Shamim AA, Viljakainen HT, Akhtaruzzaman M, Jehan AH, Khan HU, Al-Arif FA and Lamberg-Allardt C, 2010. Effect of vitamin D, calcium and multiple micronutrient supplementation on vitamin D and bone status in Bangladeshi premenopausal garment factory workers with hypovitaminosis D: a double-blinded, randomised, placebo-controlled 1-year intervention. British Journal of Nutrition, 104, 241– 247.
Itkonen ST, Skaffari E, Saaristo P, Saarnio EM, Erkkola M, Jakobsen J, Cashman KD and Lamberg-Allardt C, 2016. Effects of vitamin D2-fortified bread v. supplementation with vitamin D2 or D3 on serum 25-hydroxyvitamin D metabolites: an 8-week randomised-controlled trial in young adult Finnish women. British Journal of Nutrition, 115, 1232– 1239.
Ito M, Koyama H, Ohshige A, Maeda T, Yoshimura T and Okamura H, 1994. Prevention of preeclampsia with calcium supplementation and vitamin D3 in an antenatal protocol. International Journal of Gynaecology and Obstetrics, 47, 115– 120.
Jackson RD, LaCroix AZ, Gass M, Wallace RB, Robbins J, Lewis CE, Bassford T, Beresford SA, Black HR, Blanchette P, Bonds DE, Brunner RL, Brzyski RG, Caan B, Cauley JA, Chlebowski RT, Cummings SR, Granek I, Hays J, Heiss G, Hendrix SL, Howard BV, Hsia J, Hubbell FA, Johnson KC, Judd H, Kotchen JM, Kuller LH, Langer RD, Lasser NL, Limacher MC, Ludlam S, Manson JE, Margolis KL, McGowan J, Ockene JK, O'Sullivan MJ, Phillips L, Prentice RL, Sarto GE, Stefanick ML, Van Horn L, Wactawski-Wende J, Whitlock E, Anderson GL, Assaf AR, Barad D and Women's Health Initiative I, 2006. Calcium plus vitamin D supplementation and the risk of fractures. New England Journal of Medicine, 354, 669– 683.
Jakobsen J and Knuthsen P, 2014. Stability of vitamin D in foodstuffs during cooking. Food Chemistry, 148, 170– 175.
Jakobsen J and Saxholt E, 2009. Vitamin D metabolites in bovine milk and butter. Journal of Food Composition and Analysis, 22, 472– 478.
Janssen HC, Samson MM and Verhaar HJ, 2010. Muscle strength and mobility in vitamin D-insufficient female geriatric patients: a randomized controlled trial on vitamin D and calcium supplementation. Aging Clinical and Experimental Research, 22, 78– 84.
Jäpelt RB and Jakobsen J, 2013. Vitamin D in plants: a review of occurrence, analysis, and biosynthesis. Frontiers in Plant Science, 4, 136.
Javaid MK, Crozier SR, Harvey NC, Gale CR, Dennison EM, Boucher BJ, Arden NK, Godfrey KM and Cooper C and Princess Anne Hospital Study Group, 2006. Maternal vitamin D status during pregnancy and childhood bone mass at age 9 years: a longitudinal study. Lancet, 367, 36– 43.
Jeans PC and Stearns G, 1938. The effect of vitamin D on linear growth in infancy: II. The effect of intakes above 1,800 U.S.P. units daily. Journal of Pediatrics, 13, 730– 740.
Jensen C, Holloway L, Block G, Spiller G, Gildengorin G, Gunderson E, Butterfield G and Marcus R, 2002. Long-term effects of nutrient intervention on markers of bone remodeling and calciotropic hormones in late-postmenopausal women. American Journal of Clinical Nutrition, 75, 1114– 1120.
Johnsen MS, Grimnes G, Figenschau Y, Torjesen PA, Almas B and Jorde R, 2014. Serum free and bio-available 25-hydroxyvitamin D correlate better with bone density than serum total 25-hydroxyvitamin D. Scandinavian Journal of Clinical and Laboratory Investigation, 74, 177– 183.
Johnson JL, Mistry VV, Vukovich MD, Hogie-Lorenzen T, Hollis BW and Specker BL, 2005. Bioavailability of vitamin D from fortified process cheese and effects on vitamin D status in the elderly. Journal of Dairy Science, 88, 2295– 2301.
Jones G, 2013. Extrarenal vitamin D activation and interactions between vitamin D(2), vitamin D(3), and vitamin D analogs. Annual Review of Nutrition, 33, 23– 44.
Jones G, 2014. Vitamin D. In: AC Ross, B Caballero, RJ Cousins, KL Tucker, TR Ziegler (eds.). Modern Nutrition in Health and Disease. Wolters Kluwer Health, Lippincott Williams & Wilkins, Baltimore, MD, Philadelphia, PA, USA. pp. 278– 292.
Jones G, Strugnell SA and DeLuca HF, 1998. Current understanding of the molecular actions of vitamin D. Physiological Reviews, 78, 1193– 1231.
Jones G, Prosser DE and Kaufmann M, 2012. 25-Hydroxyvitamin D-24-hydroxylase (CYP24A1): Its important role in the degradation of vitamin D. Archives of Biochemistry and Biophysics, 523, 9– 18.
Jones G, Prosser DE and Kaufmann M, 2014. Cytochrome P450-mediated metabolism of vitamin D. Journal of Lipid Research, 55, 13– 31.
Jones KS, Assar S, Harnpanich D, Bouillon R, Lambrechts D, Prentice A and Schoenmakers I, 2014. 25(OH)D2 half-life is shorter than 25(OH)D3 half-life and is influenced by DBP concentration and genotype. Journal of Clinical Endocrinology and Metabolism, 99, 3373– 3381.
Jorde R, Sneve M, Torjesen PA, Figenschau Y, Hansen JB and Grimnes G, 2010. No significant effect on bone mineral density by high doses of vitamin D3 given to overweight subjects for one year. Nutrition Journal, 9, 1.
Jungert A, Roth HJ and Neuhäuser-Berthold M, 2012. Serum 25-hydroxyvitamin D3 and body composition in an elderly cohort from Germany: a cross-sectional study. Nutrition and Metabolism, 9, 42.
Kalkwarf HJ, Specker BL, Heubi JE, Vieira NE and Yergey AL, 1996. Intestinal calcium absorption of women during lactation and after weaning. American Journal of Clinical Nutrition, 63, 526– 531.
Kalyani RR, Stein B, Valiyil R, Manno R, Maynard JW and Crews DC, 2010. Vitamin D treatment for the prevention of falls in older adults: systematic review and meta-analysis. Journal of the American Geriatrics Society, 58, 1299– 1310.
Kamao M, Tatematsu S, Hatakeyama S, Sakaki T, Sawada N, Inouye K, Ozono K, Kubodera N, Reddy GS and Okano T, 2004. C-3 epimerization of vitamin D3 metabolites and further metabolism of C-3 epimers: 25-hydroxyvitamin D3 is metabolized to 3-epi-25-hydroxyvitamin D3 and subsequently metabolized through C-1alpha or C-24 hydroxylation. Journal of Biological Chemistry, 279, 15897– 15907.
Kärkkäinen M, Tuppurainen M, Salovaara K, Sandini L, Rikkonen T, Sirola J, Honkanen R, Jurvelin J, Alhava E and Kröger H, 2010. Effect of calcium and vitamin D supplementation on bone mineral density in women aged 65-71 years: a 3-year randomized population-based trial (OSTPRE-FPS). Osteoporosis International, 21, 2047– 2055.
Keane EM, Healy M, O'Moore R, Coakley D and Walsh JB, 1998. Vitamin D-fortified liquid milk: benefits for the elderly community-based population. Calcified Tissue International, 62, 300– 302.
Kent GN, Price RI, Gutteridge DH, Smith M, Allen JR, Bhagat CI, Barnes MP, Hickling CJ, Retallack RW and Wilson SG, 1990. Human lactation: forearm trabecular bone loss, increased bone turnover, and renal conservation of calcium and inorganic phosphate with recovery of bone mass following weaning. Journal of Bone and Mineral Research, 5, 361– 369.
Khadilkar A, Kadam N, Chiplonkar S, Fischer PR and Khadilkar V, 2012. School-based calcium-vitamin D with micronutrient supplementation enhances bone mass in underprivileged Indian premenarchal girls. Bone, 51, 1– 7.
Kift R, Berry JL, Vail A, Durkin MT, Rhodes LE and Webb AR, 2013. Lifestyle factors including less cutaneous sun exposure contribute to starkly lower vitamin D levels in U.K. South Asians compared with the white population. British Journal of Dermatology, 169, 1272– 1278.
Kim MJ, Na B, No SJ, Han HS, Jeong EH, Lee W, Han Y and Hyeun T, 2010. Nutritional status of vitamin D and the effect of vitamin D supplementation in Korean breast-fed infants. Journal of Korean Medical Science, 25, 83– 89.
Kinyamu HK, Gallagher JC, Rafferty KA and Balhorn KE, 1998. Dietary calcium and vitamin D intake in elderly women: effect on serum parathyroid hormone and vitamin D metabolites. American Journal of Clinical Nutrition, 67, 342– 348.
Kitamura K, Nakamura K, Saito T, Kobayashi R, Oshiki R, Nishiwaki T, Iwasaki M and Yoshihara A, 2013. High serum 25-hydroxyvitamin D levels do not retard postmenopausal bone loss in Japanese women: the Yokogoshi study. Archives of Osteoporosis, 8, 153.
Kitanaka S, Takeyama K, Murayama A, Sato T, Okumura K, Nogami M, Hasegawa Y, Niimi H, Yanagisawa J, Tanaka T and Kato S, 1998. Inactivating mutations in the 25-hydroxyvitamin D3 1alpha-hydroxylase gene in patients with pseudovitamin D-deficiency rickets. New England Journal of Medicine, 338, 653– 661.
Knapp G and Hartung J, 2003. Improved tests for a random effects meta-regression with a single covariate. Statistics in Medicine, 22, 2693– 2710.
Knutsen KV, Madar AA, Lagerlov P, Brekke M, Raastad T, Stene LC and Meyer HE, 2014. Does vitamin D improve muscle strength in adults? A randomized, double-blind, placebo-controlled trial among ethnic minorities in Norway. Journal of Clinical Endocrinology and Metabolism, 99, 194– 202.
Komulainen MH, Kroger H, Tuppurainen MT, Heikkinen AM, Alhava E, Honkanen R and Saarikoski S, 1998. HRT and Vit D in prevention of non-vertebral fractures in postmenopausal women; a 5 year randomized trial. Maturitas, 31, 45– 54.
Kovacs CS and Kronenberg HM, 1997. Maternal-fetal calcium and bone metabolism during pregnancy, puerperium, and lactation. Endocrine Reviews, 18, 832– 872.
Kristensen HL, Rosenqvist E and Jakobsen J, 2012. Increase of vitamin D(2) by UV-B exposure during the growth phase of white button mushroom (Agaricus bisporus). Food & Nutrition Research, 2012. 56, 7114 doi: 10.3402/fnr.v56i0.7114.
Kuan V, Martineau AR, Griffiths CJ, Hypponen E and Walton R, 2013. DHCR7 mutations linked to higher vitamin D status allowed early human migration to northern latitudes. BioMed Central Evolutionary Biology, 13, 144.
Kuhn J, Schutkowski A, Hirche F, Baur AC, Mielenz N and Stangl GI, 2015. Non-linear increase of vitamin D content in eggs from chicks treated with increasing exposure times of ultraviolet light. Journal of Steroid Biochemistry and Molecular Biology, 148, 7– 13.
Kukuljan S, Nowson CA, Sanders KM, Nicholson GC, Seibel MJ, Salmon J and Daly RM, 2011. Independent and combined effects of calcium-vitamin D3 and exercise on bone structure and strength in older men: an 18-month factorial design randomized controlled trial. Journal of Clinical Endocrinology and Metabolism, 96, 955– 963.
Kumar GT, Sachdev HS, Chellani H, Rehman AM, Singh V, Arora H and Filteau S, 2011. Effect of weekly vitamin D supplements on mortality, morbidity, and growth of low birthweight term infants in India up to age 6 months: randomised controlled trial. British Medical Journal, 342, d2975.
Laaksonen MM, Karkkainen MU, Outila TA, Rita HJ and Lamberg-Allardt CJ, 2004. Vitamin D receptor gene start codon polymorphism (FokI) is associated with forearm bone mineral density and calcaneal ultrasound in Finnish adolescent boys but not in girls. Journal of Bone and Mineral Metabolism, 22, 479– 485.
LaCroix AZ, Kotchen J, Anderson G, Brzyski R, Cauley JA, Cummings SR, Gass M, Johnson KC, Ko M, Larson J, Manson JE, Stefanick ML and Wactawski-Wende J, 2009. Calcium plus vitamin D supplementation and mortality in postmenopausal women: the Women's Health Initiative calcium-vitamin D randomized controlled trial. . Journals of Gerontology Series A, Biological Sciences and Medical Sciences, 64, 559– 567.
Lagari V, Gomez-Marin O and Levis S, 2013. The role of vitamin D in improving physical performance in the elderly. Journal of Bone and Mineral Research, 28, 2194– 2201.
Lalau JD, Jans I, el Esper N, Bouillon R and Fournier A, 1993. Calcium metabolism, plasma parathyroid hormone, and calcitriol in transient hypertension of pregnancy. American Journal of Hypertension, 6, 522– 527.
Lamberg-Allardt C, 1984. Vitamin D intake, sunlight exposure and 25-hydroxyvitamin D levels in the elderly during one year. Annals of Nutrition and Metabolism, 28, 144– 150.
Lamberg-Allardt C, Brustad M, Meyer HE and Steingrimsdottir L, 2013. Vitamin D – a systematic literature review for the 5th edition of the Nordic Nutrition Recommendations. Food & Nutrition Research, 2013. 57, 22671 http://dx.doi.org/10.3402/fnr.v57i0.22671.
Landin-Wilhelmsen K, Wilhelmsen L and Bengtsson BA, 1999. Postmenopausal osteoporosis is more related to hormonal aberrations than to lifestyle factors. Clinical Endocrinology, 51, 387– 394.
Larsen ER, Mosekilde L and Foldspang A, 2004. Vitamin D and calcium supplementation prevents osteoporotic fractures in elderly community dwelling residents: a pragmatic population-based 3-year intervention study. Journal of Bone and Mineral Research, 19, 370– 378.
Larsen ER, Mosekilde L and Foldspang A, 2005. Vitamin D and calcium supplementation prevents severe falls in elderly community-dwelling women: a pragmatic population-based 3-year intervention study. Aging Clinical and Experimental Research, 17, 125– 132.
Larsen T, Mose FH, Bech JN, Hansen AB and Pedersen EB, 2012. Effect of cholecalciferol supplementation during winter months in patients with hypertension: a randomized, placebo-controlled trial. American Journal of Hypertension, 25, 1215– 1222.
Latham NK, Anderson CS, Lee A, Bennett DA, Moseley A and Cameron ID and Fitness Collaborative Group, 2003. A randomized, controlled trial of quadriceps resistance exercise and vitamin D in frail older people: the Frailty Interventions Trial in Elderly Subjects (FITNESS). Journal of the American Geriatrics Society, 51, 291– 299.
Lau EM, Woo J, Swaminathan R, MacDonald D and Donnan SP, 1989. Plasma 25-hydroxyvitamin D concentration in patients with hip fracture in Hong Kong. Gerontology, 35, 198– 204.
Law M, Withers H, Morris J and Anderson F, 2006. Vitamin D supplementation and the prevention of fractures and falls: results of a randomised trial in elderly people in residential accommodation. Age and Ageing, 35, 482– 486.
LeBoff MS, Kohlmeier L, Hurwitz S, Franklin J, Wright J and Glowacki J, 1999. Occult vitamin D deficiency in postmenopausal US women with acute hip fracture. JAMA, 281, 1505– 1511.
Lee WH, McCabe GP, Martin BR and Weaver CM, 2011. Simple isotopic method using oral stable or radioactive tracers for estimating fractional calcium absorption in adult women. Osteoporosis International, 22, 1829– 1834.
Leffelaar ER, Vrijkotte TG and van Eijsden M, 2010. Maternal early pregnancy vitamin D status in relation to fetal and neonatal growth: results of the multi-ethnic Amsterdam Born Children and their Development cohort. British Journal of Nutrition, 104, 108– 117.
Lehmann B and Meurer M, 2010. Vitamin D metabolism. Dermatologic Therapy, 23, 2– 12.
Lehmann U, Hirche F, Stangl GI, Hinz K, Westphal S and Dierkes J, 2013. Bioavailability of vitamin D(2) and D(3) in healthy volunteers, a randomized placebo-controlled trial. Journal of Clinical Endocrinology and Metabolism, 98, 4339– 4345.
Lehtonen-Veromaa MK, Mottonen TT, Nuotio IO, Irjala KM, Leino AE and Viikari JS, 2002. Vitamin D and attainment of peak bone mass among peripubertal Finnish girls: a 3-y prospective study. American Journal of Clinical Nutrition, 76, 1446– 1453.
Lehtonen-Veromaa M, Mottonen T, Leino A, Heinonen OJ, Rautava E and Viikari J, 2008. Prospective study on food fortification with vitamin D among adolescent females in Finland: minor effects. British Journal of Nutrition, 100, 418– 423.
Lemann J Jr, Pleuss JA, Gray RW and Hoffmann RG, 1991. Potassium administration reduces and potassium deprivation increases urinary calcium excretion in healthy adults [corrected]. Kidney International, 39, 973– 983.
Lerch C and Meissner T, 2007. Interventions for the prevention of nutritional rickets in term born children. Cochrane Database of Systematic Reviews, 4, CD006164.
Lewis RD, Laing EM, Hill Gallant KM, Hall DB, McCabe GP, Hausman DB, Martin BR, Warden SJ, Peacock M and Weaver CM, 2013. A randomized trial of vitamin D(3) supplementation in children: dose-response effects on vitamin D metabolites and calcium absorption. Journal of Clinical Endocrinology and Metabolism, 98, 4816– 4825.
Linseisen J, Bechthold A, Bischoff-Ferrari HA, Hintzpeter B, Leschik-Bonnet E, Reichrath J, Stehle P, Volkert D, Wolfram G and Zittermann A, 2011. Stellungnahme. Vitamin D und Prävention ausgewählter chronischer Krankheiten. Deutschen Gesellschaft für Ernährung e. V. (DGE), 47 pp.
Lips P, 2006. Vitamin D physiology. Progress in Biophysics and Molecular Biology, 92, 4– 8.
Lips P, Hackeng WH, Jongen MJ, van Ginkel FC and Netelenbos JC, 1983. Seasonal variation in serum concentrations of parathyroid hormone in elderly people. Journal of Clinical Endocrinology and Metabolism, 57, 204– 206.
Lips P, van Ginkel FC, Jongen MJ, Rubertus F, van der Vijgh WJ and Netelenbos JC, 1987. Determinants of vitamin D status in patients with hip fracture and in elderly control subjects. American Journal of Clinical Nutrition, 46, 1005– 1010.
Lips P, Graafmans WC, Ooms ME, Bezemer PD and Bouter LM, 1996. Vitamin D supplementation and fracture incidence in elderly persons. A randomized, placebo-controlled clinical trial. Annals of Internal Medicine, 124, 400– 406.
Lips P, Binkley N, Pfeifer M, Recker R, Samanta S, Cohn DA, Chandler J, Rosenberg E and Papanicolaou DA, 2010. Once-weekly dose of 8400 IU vitamin D(3) compared with placebo: effects on neuromuscular function and tolerability in older adults with vitamin D insufficiency. American Journal of Clinical Nutrition, 91, 985– 991.
Liu J, Greenfield H, Strobel N and Fraser DR, 2013. The influence of latitude on the concentration of vitamin D3 and 25-hydroxy-vitamin D3 in Australian red meat. Food Chemistry, 140, 432– 435.
Lo CW, Paris PW, Clemens TL, Nolan J and Holick MF, 1985. Vitamin D absorption in healthy subjects and in patients with intestinal malabsorption syndromes. American Journal of Clinical Nutrition, 42, 644– 649.
Looker AC, 2013. Serum 25-hydroxyvitamin D and risk of major osteoporotic fractures in older U.S. adults. Journal of Bone and Mineral Research, 28, 997– 1006.
Looker AC and Mussolino ME, 2008. Serum 25-hydroxyvitamin D and hip fracture risk in older U.S. white adults. Journal of Bone and Mineral Research, 23, 143– 150.
Lund B, Sorensen OH and Christensen AB, 1975. 25-Hydroxycholecaliferol and fractures of the proximal. Lancet, 2, 300– 302.
Lyons RA, Johansen A, Brophy S, Newcombe RG, Phillips CJ, Lervy B, Evans R, Wareham K and Stone MD, 2007. Preventing fractures among older people living in institutional care: a pragmatic randomised double blind placebo controlled trial of vitamin D supplementation. Osteoporosis International, 18, 811– 818.
MacDonald HM, Mavroeidi A, Fraser WD, Darling AL, Black AJ, Aucott L, O'Neill F, Hart K, Berry JL, Lanham-New SA and Reid DM, 2011. Sunlight and dietary contributions to the seasonal vitamin D status of cohorts of healthy postmenopausal women living at northerly latitudes: a major cause for concern? Osteoporosis International, 22, 2461– 2472.
Macdonald HM, Wood AD, Aucott LS, Black AJ, Fraser WD, Mavroeidi A, Reid DM, Secombes KR, Simpson WG and Thies F, 2013. Hip bone loss is attenuated with 1000 IU but not 400 IU daily vitamin D3: a 1-year double-blind RCT in postmenopausal women. Journal of Bone and Mineral Research, 28, 2202– 2213.
MacLaughlin J and Holick MF, 1985. Aging decreases the capacity of human skin to produce vitamin D3. Journal of Clinical Investigation, 76, 1536– 1538.
MacLennan WJ and Hamilton JC, 1977. Vitamin D supplements and 25-hydroxy vitamin D concentrations in the elderly. British Medical Journal, 2, 859– 861.
Madsen KH, Rasmussen LB, Andersen R, Molgaard C, Jakobsen J, Bjerrum PJ, Andersen EW, Mejborn H and Tetens I, 2013. Randomized controlled trial of the effects of vitamin D-fortified milk and bread on serum 25-hydroxyvitamin D concentrations in families in Denmark during winter: the VitmaD study. American Journal of Clinical Nutrition, 98, 374– 382.
Mahon P, Harvey N, Crozier S, Inskip H, Robinson S, Arden N, Swaminathan R, Cooper C and Godfrey K and Group SWSS, 2010. Low maternal vitamin D status and fetal bone development: cohort study. Journal of Bone and Mineral Research, 25, 14– 19.
Majid Molla A, Badawi MH, al-Yaish S, Sharma P, el-Salam RS and Molla AM, 2000. Risk factors for nutritional rickets among children in Kuwait. Pediatrics International, 42, 280– 284.
Makin HL, Seamark DA and Trafford DJ, 1983. Vitamin D and its metabolites in human breast milk. Archives of Disease in Childhood, 58, 750– 753.
Mallet E, Gugi B, Brunelle P, Henocq A, Basuyau JP and Lemeur H, 1986. Vitamin D supplementation in pregnancy: a controlled trial of two methods. Obstetrics and Gynecology, 68, 300– 304.
Malmberg P, Karlsson T, Svensson H, Lonn M, Carlsson NG, Sandberg AS, Jennische E, Osmancevic A and Holmang A, 2014. A new approach to measuring vitamin D in human adipose tissue using time-of-flight secondary ion mass spectrometry: a pilot study. Journal of Photochemistry and Photobiology. B, Biology, 138, 295– 301.
Markestad T, 1983. Effect of season and vitamin D supplementation on plasma concentrations of 25-hydroxyvitamin D in Norwegian infants. Acta Paediatrica Scandinavica, 72, 817– 821.
Markestad T and Elzouki AY, 1991. Vitamin D-deficiency rickets in northern Europe and Libya. In: F Glorieux (ed.). Rickets. Nestlé Nutrition Workshop Series, 21. Raven Press, New York, NY, USA. pp. 203– 213.
Markestad T, Halvorsen S, Halvorsen KS, Aksnes L and Aarskog D, 1984. Plasma-concentrations of vitamin-D metabolites before and during treatment of vitamin-D deficiency rickets in children. Acta Paediatrica Scandinavica, 73, 225– 231.
Markestad T, Ulstein M, Aksnes L and Aarskog D, 1986. Serum concentrations of vitamin D metabolites in vitamin D supplemented pregnant women. A longitudinal study. Acta Obstetricia et Gynecologica Scandinavica, 65, 63– 67.
Marwaha RK, Tandon N, Reddy DR, Aggarwal R, Singh R, Sawhney RC, Saluja B, Ganie MA and Singh S, 2005. Vitamin D and bone mineral density status of healthy schoolchildren in northern India. American Journal of Clinical Nutrition, 82, 477– 482.
Marya RK, Rathee S and Manrow M, 1987. Effect of calcium and vitamin D supplementation on toxaemia of pregnancy. Gynecologic and Obstetric Investigation, 24, 38– 42.
Marya RK, Rathee S, Dua V and Sangwan K, 1988. Effect of vitamin D supplementation during pregnancy on foetal growth. Indian Journal of Medical Research, 88, 488– 492.
Mattila PH, Piironen VI, Uusi-Rauva EJ and Koivistoinen PE, 1993. Determination of 25-hydroxycholecalciferol content in egg yolk by HPLC. Journal of Food Composition and Analysis, 6, 250– 255.
Mattila PH, Piironen VI, Uusi-Rauva EJ and Koivistoinen PE, 1995. Contents of cholecalciferol, ergocalciferol, and their 25-hydroxylated metabolites in milk products and raw meat and liver as determined by HPLC. Journal of Agricultural and Food Chemistry, 43, 2394– 2399.
Mattila P, Lehikoinen K, Kiiskinen T and Piironen V, 1999. Cholecalciferol and 25-hydroxycholecalciferol content of chicken egg yolk as affected by the cholecalciferol content of feed. Journal of Agricultural and Food Chemistry, 47, 4089– 4092.
Mattila PH, Valkonen E and Valaja J, 2011. Effect of different vitamin D supplementations in poultry feed on vitamin D content of eggs and chicken meat. Journal of Agricultural and Food Chemistry, 59, 8298– 8303.
Mawer EB, Backhouse J, Holman CA, Lumb GA and Stanbury SW, 1972. The distribution and storage of vitamin D and its metabolites in human tissues. Clinical Science, 43, 413– 431.
Maxwell JP and Miles LM, 1925. Osteomalacia in China. Proceedings of the Royal Society of Medicine, 18, 48– 66.
Maxwell JD, Ang L, Brooke OG and Brown IR, 1981. Vitamin D supplements enhance weight gain and nutritional status in pregnant Asians. British Journal of Obstetrics and Gynaecology, 88, 987– 991.
McGrath JJ, Saha S, Burne TH and Eyles DW, 2010. A systematic review of the association between common single nucleotide polymorphisms and 25-hydroxyvitamin D concentrations. Journal of Steroid Biochemistry and Molecular Biology, 121, 471– 477.
McKenzie RL, Liley JB and Bjorn LO, 2009. UV radiation: balancing risks and benefits. Photochemistry and Photobiology, 85, 88– 98.
Meier C, Woitge HW, Witte K, Lemmer B and Seibel MJ, 2004. Supplementation with oral vitamin D3 and calcium during winter prevents seasonal bone loss: a randomized controlled open-label prospective trial. Journal of Bone and Mineral Research, 19, 1221– 1230.
Melhus H, Snellman G, Gedeborg R, Byberg L, Berglund L, Mallmin H, Hellman P, Blomhoff R, Hagstrom E, Arnlov J and Michaelsson K, 2010. Plasma 25-hydroxyvitamin D levels and fracture risk in a community-based cohort of elderly men in Sweden. Journal of Clinical Endocrinology and Metabolism, 95, 2637– 2645.
Melin A, Wilske J, Ringertz H and Saaf M, 2001. Seasonal variations in serum levels of 25-hydroxyvitamin D and parathyroid hormone but no detectable change in femoral neck bone density in an older population with regular outdoor exposure. Journal of the American Geriatrics Society, 49, 1190– 1196.
Menant JC, Close JC, Delbaere K, Sturnieks DL, Trollor J, Sachdev PS, Brodaty H and Lord SR, 2012. Relationships between serum vitamin D levels, neuromuscular and neuropsychological function and falls in older men and women. Osteoporosis International, 23, 981– 989.
Michael YL, Lin JS, Whitlock EP, Gold R, Fu R, O'Connor EA, Zuber SP, Beil TL and Lutz KW, 2010. Interventions to prevent falls in older adults: an updated systematic reviewEvidence Synthesis No. 80. AHRQ Publication No. 11–05150-EF-1. Agency for Healthcare Research and Quality, Rockville, USA. 392 pp.
Michael YL, Smit E, Seguin R, Curb JD, Phillips LS and Manson JE, 2011. Serum 25-hydroxyvitamin D and physical performance in postmenopausal women. Journal of Women's Health, 20, 1603– 1608.
Mocanu V, Stitt PA, Costan AR, Voroniuc O, Zbranca E, Luca V and Vieth R, 2009. Long-term effects of giving nursing home residents bread fortified with 125 microg (5000 IU) vitamin D(3) per daily serving. American Journal of Clinical Nutrition, 89, 1132– 1137.
Molgaard C, Larnkjaer A, Cashman KD, Lamberg-Allardt C, Jakobsen J and Michaelsen KF, 2010. Does vitamin D supplementation of healthy Danish Caucasian girls affect bone turnover and bone mineralization? Bone, 46, 432– 439.
Moller UK, Streym S, Heickendorff L, Mosekilde L and Rejnmark L, 2012. Effects of 25OHD concentrations on chances of pregnancy and pregnancy outcomes: a cohort study in healthy Danish women. European Journal of Clinical Nutrition, 66, 862– 868.
Morley R, Carlin JB, Pasco JA and Wark JD, 2006. Maternal 25-hydroxyvitamin D and parathyroid hormone concentrations and offspring birth size. Journal of Clinical Endocrinology and Metabolism, 91, 906– 912.
Moschonis G and Manios Y, 2006. Skeletal site-dependent response of bone mineral density and quantitative ultrasound parameters following a 12-month dietary intervention using dairy products fortified with calcium and vitamin D: the Postmenopausal Health Study. British Journal of Nutrition, 96, 1140– 1148.
Muir SW and Montero-Odasso M, 2011. Effect of vitamin D supplementation on muscle strength, gait and balance in older adults: a systematic review and meta-analysis. Journal of the American Geriatrics Society, 59, 2291– 2300.
Mundy GR and Guise TA, 1999. Hormonal control of calcium homeostasis. Clinical Chemistry, 45, 1347– 1352.
Murad MH, Elamin KB, Abu Elnour NO, Elamin MB, Alkatib AA, Fatourechi MM, Almandoz JP, Mullan RJ, Lane MA, Liu H, Erwin PJ, Hensrud DD and Montori VM, 2011. Clinical review: the effect of vitamin D on falls: a systematic review and meta-analysis. Journal of Clinical Endocrinology and Metabolism, 96, 2997– 3006.
Nakamura K, Saito T, Oyama M, Oshiki R, Kobayashi R, Nishiwaki T, Nashimoto M and Tsuchiya Y, 2011. Vitamin D sufficiency is associated with low incidence of limb and vertebral fractures in community-dwelling elderly Japanese women: the Muramatsu Study. Osteoporosis International, 22, 97– 103.
Namgung R, Tsang RC, Lee C, Han DG, Ho ML and Sierra RI, 1998. Low total body bone mineral content and high bone resorption in Korean winter-born versus summer-born newborn infants. Journal of Pediatrics, 132, 421– 425.
Need AG, Horowitz M, Morris HA and Nordin BC, 2000. Vitamin D status: effects on parathyroid hormone and 1, 25-dihydroxyvitamin D in postmenopausal women. American Journal of Clinical Nutrition, 71, 1577– 1581.
Need AG, O'Loughlin PD, Morris HA, Coates PS, Horowitz M and Nordin BE, 2008. Vitamin D metabolites and calcium absorption in severe vitamin D deficiency. Journal of Bone and Mineral Research, 23, 1859– 1863.
Nelson ML, Blum JM, Hollis BW, Rosen C and Sullivan SS, 2009. Supplements of 20 microg/d cholecalciferol optimized serum 25-hydroxyvitamin D concentrations in 80% of premenopausal women in winter. Journal of Nutrition, 139, 540– 546.
Newberry SJ, Chung M, Shekelle PG, Booth MS, Liu JL, Maher AR, Motala A, Cui M, Perry T, Shanman R and Balk EM, 2014. Vitamin D and Calcium: A Systematic Review of Health Outcomes (update)Evidence Report/Technology Assessment No. 217. Prepared by the Southern California Evidence-based Practice Center under Contract No. 290–2012-00006-I. AHRQ Publication No. 14–E004-EF. Agency for Healthcare Research and Quality, Rockville, USA. 929pp .
Nieves JW, Cosman F, Grubert E, Ambrose B, Ralston SH and Lindsay R, 2012. Skeletal effects of vitamin D supplementation in postmenopausal black women. Calcified Tissue International, 91, 316– 324.
Nissen J, Rasmussen LB, Ravn-Haren G, Andersen EW, Hansen B, Andersen R, Mejborn H, Madsen KH and Vogel U, 2014. Common variants in CYP2R1 and GC genes predict vitamin D concentrations in healthy Danish children and adults. PLoS ONE, 9, e89907.
NNR (Nordic Nutrition Recommendations), 2004. Integrating Nutrition and Physical Activity. Nordic Council of Ministers, Copenhagen, Denmark. 436 pp.
Nordic Council of Ministers, 2014. Nordic Nutrition Recommendations 2012. Integrating nutrition and physical activity. Nordic Council of Ministers, Copenhagen, Denmark. 627 pp.
Norman AW, 2008. From vitamin D to hormone D: fundamentals of the vitamin D endocrine system essential for good health. American Journal of Clinical Nutrition, 88, 491S– 499S.
Norman AW, 2012. The history of the discovery of vitamin D and its daughter steroid hormone. Annals of Nutrition and Metabolism, 61, 199– 206.
O'Connor E, Molgaard C, Michaelsen KF, Jakobsen J and Cashman KD, 2010. Vitamin D-vitamin K interaction: effect of vitamin D supplementation on serum percentage undercarboxylated osteocalcin, a sensitive measure of vitamin K status, in Danish girls. British Journal of Nutrition, 104, 1091– 1095.
Oginni LM, Worsfold M, Oyelami OA, Sharp CA, Powell DE and Davie MW, 1996. Etiology of rickets in Nigerian children. Journal of Pediatrics, 128, 692– 694.
Okonofua F, Menon RK, Houlder S, Thomas M, Robinson D, O'Brien S and Dandona P, 1987. Calcium, vitamin D and parathyroid hormone relationships in pregnant Caucasian and Asian women and their neonates. Annals of Clinical Biochemistry, 24(Pt 1), 22– 28.
Olafsdottir AS, Wagner KH, Thorsdottir I and Elmadfa I, 2001. Fat-soluble vitamins in the maternal diet, influence of cod liver oil supplementation and impact of the maternal diet on human milk composition. Annals of Nutrition and Metabolism, 45, 265– 272.
Onder G, Capoluongo E, Danese P, Settanni S, Russo A, Concolino P, Bernabei R and Landi F, 2008. Vitamin D receptor polymorphisms and falls among older adults living in the community: results from the ilSIRENTE study. Journal of Bone and Mineral Research, 23, 1031– 1036.
Ooms ME, Roos JC, Bezemer PD, van der Vijgh WJ, Bouter LM and Lips P, 1995. Prevention of bone loss by vitamin D supplementation in elderly women: a randomized double-blind trial. Journal of Clinical Endocrinology and Metabolism, 80, 1052– 1058.
Ovesen L, Brot C and Jakobsen J, 2003. Food contents and biological activity of 25-hydroxyvitamin D: a vitamin D metabolite to be reckoned with? Annals of Nutrition and Metabolism, 47, 107– 113.
Parikh SJ, Edelman M, Uwaifo GI, Freedman RJ, Semega-Janneh M, Reynolds J and Yanovski JA, 2004. The relationship between obesity and serum 1,25-dihydroxy vitamin D concentrations in healthy adults. Journal of Clinical Endocrinology and Metabolism, 89, 1196– 1199.
Park MJ, Namgung R, Kim DH and Tsang RC, 1998. Bone mineral content is not reduced despite low vitamin D status in breast milk-fed infants versus cow's milk based formula-fed infants. Journal of Pediatrics, 132, 641– 645.
Park CY, Hill KM, Elble AE, Martin BR, DiMeglio LA, Peacock M, McCabe GP and Weaver CM, 2010. Daily supplementation with 25 mug cholecalciferol does not increase calcium absorption or skeletal retention in adolescent girls with low serum 25-hydroxyvitamin D. Journal of Nutrition, 140, 2139– 2144.
Patel R, Collins D, Bullock S, Swaminathan R, Blake GM and Fogelman I, 2001. The effect of season and vitamin D supplementation on bone mineral density in healthy women: a double-masked crossover study. Osteoporosis International, 12, 319– 325.
Pekkarinen T, Valimaki MJ, Velimaki VV, Aarum S, Turpeinen U, Hamalainen E and Loyttyniemi E, 2010. The same annual dose of 292000 IU of vitamin D3 (cholecalciferol) on either daily or four monthly basis for elderly women: 1-year comparative study of the effects on serum 25(OH)D3 concentrations and renal function. Clinical Endocrinology, 72, 455– 461.
Pereira GR and Zucker AH, 1986. Nutritional deficiencies in the neonate. Clinics in Perinatology, 13, 175– 189.
Pettifor JM and Daniels ED, 1997. Vitamin D deficiency and nutritional rickets in children. In: D Feldman, FH Glorieux and JW Pike (eds.). Vitamin D. Academic Press, New York, NY, USA. pp. 663– 678.
Pfeifer M, Begerow B, Minne HW, Abrams C, Nachtigall D and Hansen C, 2000. Effects of a short-term vitamin D and calcium supplementation on body sway and secondary hyperparathyroidism in elderly women. Journal of Bone and Mineral Research, 15, 1113– 1118.
Pfeifer M, Begerow B, Minne HW, Nachtigall D and Hansen C, 2001. Effects of a short-term vitamin D(3) and calcium supplementation on blood pressure and parathyroid hormone levels in elderly women. Journal of Clinical Endocrinology and Metabolism, 86, 1633– 1637.
Pfeifer M, Begerow B, Minne HW, Suppan K, Fahrleitner-Pammer A and Dobnig H, 2009. Effects of a long-term vitamin D and calcium supplementation on falls and parameters of muscle function in community-dwelling older individuals. Osteoporosis International, 20, 315– 322.
Pirotta S, Kidgell DJ and Daly RM, 2015. Effects of vitamin D supplementation on neuroplasticity in older adults: a double-blinded, placebo-controlled randomised trial. Osteoporosis International, 26, 131– 140.
Porojnicu AC, Bruland OS, Aksnes L, Grant WB and Moan J, 2008. Sun beds and cod liver oil as vitamin D sources. Journal of Photochemistry and Photobiology B: Biology, 91, 125– 131.
Porthouse J, Cockayne S, King C, Saxon L, Steele E, Aspray T, Baverstock M, Birks Y, Dumville J, Francis R, Iglesias C, Puffer S, Sutcliffe A, Watt I and Torgerson DJ, 2005. Randomised controlled trial of calcium and supplementation with cholecalciferol (vitamin D3) for prevention of fractures in primary care. British Medical Journal, 330, 1003.
Poskitt EM, Cole TJ and Lawson DE, 1979. Diet, sunlight, and 25-hydroxy vitamin D in healthy children and adults. British Medical Journal, 1, 221– 223.
Powe CE, Seely EW, Rana S, Bhan I, Ecker J, Karumanchi SA and Thadhani R, 2010. First trimester vitamin D, vitamin D binding protein, and subsequent preeclampsia. Hypertension, 56, 758– 763.
Powe CE, Evans MK, Wenger J, Zonderman AB, Berg AH, Nalls M, Tamez H, Zhang D, Bhan I, Karumanchi SA, Powe NR and Thadhani R, 2013. Vitamin D-binding protein and vitamin D status of black Americans and white Americans. New England Journal of Medicine, 369, 1991– 2000.
Preece MA, Tomlinson S, Ribot CA, Pietrek J, Korn HT, Davies DM, Ford JA, Dunnigan MG and O'Riordan JL, 1975. Studies of vitamin D deficiency in man. Quarterly Journal of Medicine, 44, 575– 589.
Prentice A, 1998. Calcium requirements of breast-feeding mothers. Nutrition Reviews, 56, 124– 127.
Prentice A, Yan L, Jarjou LM, Dibba B, Laskey MA, Stirling DM and Fairweather-Tait S, 1997. Vitamin D status does not influence the breast-milk calcium concentration of lactating mothers accustomed to a low calcium intake. Acta Paediatrica, 86, 1006– 1008.
Prentice A, Jarjou LM, Goldberg GR, Bennett J, Cole TJ and Schoenmakers I, 2009. Maternal plasma 25-hydroxyvitamin D concentration and birthweight, growth and bone mineral accretion of Gambian infants. Acta Paediatrica, 98, 1360– 1362.
Prentice RL, Pettinger MB, Jackson RD, Wactawski-Wende J, Lacroix AZ, Anderson GL, Chlebowski RT, Manson JE, Van Horn L, Vitolins MZ, Datta M, LeBlanc ES, Cauley JA and Rossouw JE, 2013. Health risks and benefits from calcium and vitamin D supplementation: Women's Health Initiative clinical trial and cohort study. Osteoporosis International, 24, 567– 580.
Priemel M, von Domarus C, Klatte TO, Kessler S, Schlie J, Meier S, Proksch N, Pastor F, Netter C, Streichert T, Puschel K and Amling M, 2010. Bone mineralization defects and vitamin D deficiency: histomorphometric analysis of iliac crest bone biopsies and circulating 25-hydroxyvitamin D in 675 patients. Journal of Bone and Mineral Research, 25, 305– 312.
Prince RL, Austin N, Devine A, Dick IM, Bruce D and Zhu K, 2008. Effects of ergocalciferol added to calcium on the risk of falls in elderly high-risk women. Archives of Internal Medicine, 168, 103– 108.
del Puente A, Esposito A, Savastano S, Carpinelli A, Postiglione L and Oriente P, 2002. Dietary calcium intake and serum vitamin D are major determinants of bone mass variations in women. A longitudinal study. Aging Clinical and Experimental Research, 14, 382– 388.
Punnonen R, Salmi J, Tuimala R, Jarvinen M and Pystynen P, 1986. Vitamin D deficiency in women with femoral neck fracture. Maturitas, 8, 291– 295.
Reboul E, 2015. Intestinal absorption of vitamin D: from the meal to the enterocyte. Food and Function, 6, 356– 362.
Reid IR, Bolland MJ and Grey A, 2014. Effects of vitamin D supplements on bone mineral density: a systematic review and meta-analysis. Lancet, 383, 146– 155.
Reif S, Katzir Y, Eisenberg Z and Weisman Y, 1988. Serum 25-hydroxyvitamin D levels in congenital craniotabes. Acta Paediatrica Scandinavica, 77, 167– 168.
Rejnmark L, Vestergaard P, Brot C and Mosekilde L, 2011. Increased fracture risk in normocalcemic postmenopausal women with high parathyroid hormone levels: a 16-year follow-up study. Calcified Tissue International, 88, 238– 245.
Rich-Edwards JW, Ganmaa D, Kleinman K, Sumberzul N, Holick MF, Lkhagvasuren T, Dulguun B, Burke A and Frazier AL, 2011. Randomized trial of fortified milk and supplements to raise 25-hydroxyvitamin D concentrations in schoolchildren in Mongolia. American Journal of Clinical Nutrition, 94, 578– 584.
Robinson CJ, Alanis MC, Wagner CL, Hollis BW and Johnson DD, 2010. Plasma 25-hydroxyvitamin D levels in early-onset severe preeclampsia. American Journal of Obstetrics and Gynecology, 203, 366. e361-366.
Robinson CJ, Wagner CL, Hollis BW, Baatz JE and Johnson DD, 2011. Maternal vitamin D and fetal growth in early-onset severe preeclampsia. American Journal of Obstetrics and Gynecology, 204, 556 e551-554.
Robinson-Cohen C, Katz R, Hoofnagle AN, Cauley JA, Furberg CD, Robbins JA, Chen Z, Siscovick DS, de Boer IH and Kestenbaum B, 2011. Mineral metabolism markers and the long-term risk of hip fracture: the cardiovascular health study. Journal of Clinical Endocrinology and Metabolism, 96, 2186– 2193.
Rosen CJ, Morrison A, Zhou H, Storm D, Hunter SJ, Musgrave K, Chen T, Wei W and Holick MF, 1994. Elderly women in northern New England exhibit seasonal changes in bone mineral density and calciotropic hormones. Bone and Mineral, 25, 83– 92.
Rothberg AD, Pettifor JM, Cohen DF, Sonnendecker EW and Ross FP, 1982. Maternal-infant vitamin D relationships during breast-feeding. Journal of Pediatrics, 101, 500– 503.
Rouzi AA, Al-Sibiani SA, Al-Senani NS, Radaddi RM and Ardawi MS, 2012. Independent predictors of all osteoporosis-related fractures among healthy Saudi postmenopausal women: the CEOR Study. Bone, 50, 713– 722.
Saadi HF, Dawodu A, Afandi BO, Zayed R, Benedict S and Nagelkerke N, 2007. Efficacy of daily and monthly high-dose calciferol in vitamin D-deficient nulliparous and lactating women. American Journal of Clinical Nutrition, 85, 1565– 1571.
SACN (Scientific Advisory Committee on Nutrition), 2016. Vitamin D and Health report. 304 pp.
Sai AJ, Walters RW, Fang X and Gallagher JC, 2011. Relationship between vitamin D, parathyroid hormone, and bone health. Journal of Clinical Endocrinology and Metabolism, 96, E436– E446.
Salama MM and El-Sakka AS, 2010. Hypocalcemic seizures in breastfed infants with rickets secondary to severe maternal vitamin D deficiency. Pakistan Journal of Biological Sciences, 13, 437– 442.
Salovaara K, Tuppurainen M, Karkkainen M, Rikkonen T, Sandini L, Sirola J, Honkanen R, Alhava E and Kroger H, 2010. Effect of vitamin D(3) and calcium on fracture risk in 65- to 71-year-old women: a population-based 3-year randomized, controlled trial–the OSTPRE-FPS. Journal of Bone and Mineral Research, 25, 1487– 1495.
Sanders KM, Stuart AL, Williamson EJ, Simpson JA, Kotowicz MA, Young D and Nicholson GC, 2010. Annual high-dose oral vitamin D and falls and fractures in older women: a randomized controlled trial. JAMA, 303, 1815– 1822.
Saneei P, Salehi-Abargouei A and Esmaillzadeh A, 2013. Serum 25-hydroxy vitamin D levels in relation to body mass index: a systematic review and meta-analysis. Obesity Reviews, 14, 393– 404.
Sanwalka N, Khadilkar A, Chiplonkar S, Khatod K, Phadke N and Khadilkar V, 2013. Vitamin D receptor gene polymorphisms and bone mass indices in post-menarchal Indian adolescent girls. Journal of Bone and Mineral Metabolism, 31, 108– 115.
Sayers A, Fraser WD, Lawlor DA and Tobias JH, 2012. 25-Hydroxyvitamin-D3 levels are positively related to subsequent cortical bone development in childhood: findings from a large prospective cohort study. Osteoporosis International, 23, 2117– 2128.
SCF (Scientific Committe for Food), 1993. Nutrient and energy intakes for the European Community. Reports of the Scientific Committee for Food, 31st Series. Food - Science and Technique, European Commission, Luxembourg. 248 pp.
SCF (Scientific Committee on Food), 2002a. Opinion on the Tolerable Upper Intake Level of vitamin D. SCF/CS/NUT/UPPLEV/38, Final. 35 pp.
SCF (Scientific Committee on Food), 2002b. Opinion on the Tolerable Upper Intake Level of preformed vitamin A (retinol and retinyl esters). SCF/CS/NUT/UPPLEV/24, Final. 16 pp.
Schaafsma A, van Doormaal JJ, Muskiet FA, Hofstede GJ, Pakan I and van der Veer E, 2002. Positive effects of a chicken eggshell powder-enriched vitamin-mineral supplement on femoral neck bone mineral density in healthy late post-menopausal Dutch women. British Journal of Nutrition, 87, 267– 275.
Schmid A and Walther B, 2013. Natural vitamin D content in animal products. Advances in Nutrition, 4, 453– 462.
Schmidt K and Zirkler S, 2011. [Dietary efficacy of a micronutrient combination in patients with recurrent upper respiratory tract infections. Results of a placebo-controlled double-blind study]. MMW Fortschritte der Medizin, 153 (Suppl 3), 83– 89.
Scholl TO and Chen X, 2009. Vitamin D intake during pregnancy: association with maternal characteristics and infant birth weight. Early Human Development, 85, 231– 234.
Scholl TO, Chen X and Stein TP, 2013. Vitamin D, secondary hyperparathyroidism, and preeclampsia. American Journal of Clinical Nutrition, 98, 787– 793.
van Schoor NM, Visser M, Pluijm SM, Kuchuk N, Smit JH and Lips P, 2008. Vitamin D deficiency as a risk factor for osteoporotic fractures. Bone, 42, 260– 266.
Schou AJ, Heuck C and Wolthers OD, 2003. Vitamin D supplementation to healthy children does not affect serum osteocalcin or markers of type I collagen turnover. Acta Paediatrica, 92, 797– 801.
Scott D, Blizzard L, Fell J, Ding C, Winzenberg T and Jones G, 2010. A prospective study of the associations between 25-hydroxy-vitamin D, sarcopenia progression and physical activity in older adults. Clinical Endocrinology, 73, 581– 587.
Seamans KM and Cashman KD, 2009. Existing and potentially novel functional markers of vitamin D status: a systematic review. American Journal of Clinical Nutrition, 89, 1997S– 2008S.
Sebastian A, Hernandez RE, Portale AA, Colman J, Tatsuno J and Morris RC Jr, 1990. Dietary potassium influences kidney maintenance of serum phosphorus concentration. Kidney International, 37, 1341– 1349.
Seckmeyer G, Schrempf M, Wieczorek A, Riechelmann S, Graw K, Seckmeyer S and Zankl M, 2013. A novel method to calculate solar UV exposure relevant to vitamin D production in humans. Photochemistry and Photobiology, 89, 974– 983.
Seely EW, Wood RJ, Brown EM and Graves SW, 1992. Lower serum ionized calcium and abnormal calciotropic hormone levels in preeclampsia. Journal of Clinical Endocrinology and Metabolism, 74, 1436– 1440.
Sem SW, Sjoen RJ, Trygg K and Pedersen JI, 1987. Vitamin D status of two groups of elderly in Oslo: living in old people's homes and living in own homes. Comprehensive Gerontology. Section A, Clinical and Laboratory Sciences, 1, 126– 130.
Shand AW, Nassar N, Von Dadelszen P, Innis SM and Green TJ, 2010. Maternal vitamin D status in pregnancy and adverse pregnancy outcomes in a group at high risk for pre-eclampsia. British Journal of Obstetrics and Gynaecology, 117, 1593– 1598.
Shapses SA, Sukumar D, Schneider SH, Schlussel Y, Brolin RE and Taich L, 2012. Hormonal and dietary influences on true fractional calcium absorption in women: role of obesity. Osteoporosis International, 23, 2607– 2614.
Shapses SA, Sukumar D, Schneider SH, Schlussel Y, Sherrell RM, Field MP and Ambia-Sobhan H, 2013. Vitamin D supplementation and calcium absorption during caloric restriction: a randomized double-blind trial. American Journal of Clinical Nutrition, 97, 637– 645.
Shearer MJ, 1997. The roles of vitamins D and K in bone health and osteoporosis prevention. Proceedings of the Nutrition Society, 56, 915– 937.
Signaleringscommissie Kanker van KWF Kankerbestrijding, 2010. De relatie tussen kanker, zonnestraling en vitamine D. Signaleringscommissie Kanker van KWF Kankerbestrijding, Amsterdam. 172 pp.
Signorello LB, Shi J, Cai Q, Zheng W, Williams SM, Long J, Cohen SS, Li G, Hollis BW, Smith JR and Blot WJ, 2011. Common variation in vitamin D pathway genes predicts circulating 25-hydroxyvitamin D Levels among African Americans. PLoS ONE, 6, e28623.
Silver J, Landau H, Bab I, Shvil Y, Friedlaender MM, Rubinger D and Popovtzer MM, 1985. Vitamin D-dependent rickets types I and II. Diagnosis and response to therapy. Israel Journal of Medical Sciences, 21, 53– 56.
Smith H, Anderson F, Raphael H, Maslin P, Crozier S and Cooper C, 2007. Effect of annual intramuscular vitamin D on fracture risk in elderly men and women–a population-based, randomized, double-blind, placebo-controlled trial. Rheumatology, 46, 1852– 1857.
Smith SM, Gardner KK, Locke J and Zwart SR, 2009. Vitamin D supplementation during Antarctic winter. American Journal of Clinical Nutrition, 89, 1092– 1098.
Snellman G, Melhus H, Gedeborg R, Byberg L, Berglund L, Wernroth L and Michaelsson K, 2010. Determining vitamin D status: a comparison between commercially available assays. PLoS ONE, 5, e11555.
Snellman G, Byberg L, Lemming EW, Melhus H, Gedeborg R, Mallmin H, Wolk A and Michaelsson K, 2014. Long-term dietary vitamin D intake and risk of fracture and osteoporosis: a longitudinal cohort study of Swedish middle-aged and elderly women. Journal of Clinical Endocrinology and Metabolism, 99, 781– 790.
Snijder MB, van Schoor NM, Pluijm SM, van Dam RM, Visser M and Lips P, 2006. Vitamin D status in relation to one-year risk of recurrent falling in older men and women. Journal of Clinical Endocrinology and Metabolism, 91, 2980– 2985.
Sorva A, Valimaki M, Risteli J, Risteli L, Elfving S, Takkunen H and Tilvis R, 1994. Serum ionized calcium, intact PTH and novel markers of bone turnover in bedridden elderly patients. European Journal of Clinical Investigation, 24, 806– 812.
Sowers MF, Hollis BW, Shapiro B, Randolph J, Janney CA, Zhang D, Schork A, Crutchfield M, Stanczyk F and Russell-Aulet M, 1996. Elevated parathyroid hormone-related peptide associated with lactation and bone density loss. JAMA, 276, 549– 554.
Sowers M, Zhang D, Hollis BW, Shapiro B, Janney CA, Crutchfield M, Schork MA, Stanczyk F and Randolph J, 1998. Role of calciotrophic hormones in calcium mobilization of lactation. American Journal of Clinical Nutrition, 67, 284– 291.
Specker BL, 1994. Do North American women need supplemental vitamin D during pregnancy or lactation? American Journal of Clinical Nutrition, 59, 484S– 490S; discussion 490S-491S.
Specker BL, Tsang RC and Hollis BW, 1985. Effect of race and diet on human-milk vitamin D and 25-hydroxyvitamin D. American Journal of Diseases of Children, 139, 1134– 1137.
Specker BL, Ho ML, Oestreich A, Yin TA, Shui QM, Chen XC and Tsang RC, 1992. Prospective study of vitamin D supplementation and rickets in China. Journal of Pediatrics, 120, 733– 739.
Spiro A and Buttriss JL, 2014. Vitamin D: an overview of vitamin D status and intake in Europe. Nutrition Bulletin, 39, 322– 350.
Steingrimsdottir L, Halldorsson TI, Siggeirsdottir K, Cotch MF, Einarsdottir BO, Eiriksdottir G, Sigurdsson S, Launer LJ, Harris TB, Gudnason V and Sigurdsson G, 2014. Hip fractures and bone mineral density in the elderly–importance of serum 25-hydroxyvitamin D. PLoS ONE, 9, e91122.
Sterne JA and Egger M, 2001. Funnel plots for detecting bias in meta-analysis: guidelines on choice of axis. Journal of Clinical Epidemiology, 54, 1046– 1055.
Stewart G, 2009. Multiple sclerosis and vitamin D: don't (yet) blame it on the sunshine. Brain, 132, 1126– 1127.
Stockton KA, Mengersen K, Paratz JD, Kandiah D and Bennell KL, 2011. Effect of vitamin D supplementation on muscle strength: a systematic review and meta-analysis. Osteoporosis International, 22, 859– 871.
Stone K, Bauer DC, Black DM, Sklarin P, Ensrud KE and Cummings SR, 1998. Hormonal predictors of bone loss in elderly women: a prospective study. The Study of Osteoporotic Fractures Research Group. Journal of Bone and Mineral Research, 13, 1167– 1174.
Storm D, Eslin R, Porter ES, Musgrave K, Vereault D, Patton C, Kessenich C, Mohan S, Chen T, Holick MF and Rosen CJ, 1998. Calcium supplementation prevents seasonal bone loss and changes in biochemical markers of bone turnover in elderly New England women: a randomized placebo-controlled trial. Journal of Clinical Endocrinology and Metabolism, 83, 3817– 3825.
við Streym S, Hojskov CS, Moller UK, Heickendorff L, Vestergaard P, Mosekilde L and Rejnmark L, 2016. Vitamin D content in human breast milk: a 9-mo follow-up study. American Journal of Clinical Nutrition, 103, 107– 114.
Sullivan SS, Rosen CJ, Halteman WA, Chen TC and Holick MF, 2005. Adolescent girls in Maine are at risk for vitamin D insufficiency. Journal of the American Dietetic Association, 105, 971– 974.
Tai SS, Bedner M and Phinney KW, 2010. Development of a candidate reference measurement procedure for the determination of 25-hydroxyvitamin D3 and 25-hydroxyvitamin D2 in human serum using isotope-dilution liquid chromatography-tandem mass spectrometry. Analytical Chemistry, 82, 1942– 1948.
Takeda E, Yamamoto H, Taketani Y and Miyamoto K, 1997. Vitamin D-dependent rickets type I and type II. Acta Paediatrica Japonica, 39, 508– 513.
Takeuchi A, Okano T, Tsugawa N, Tasaka Y, Kobayashi T, Kodama S and Matsuo T, 1989. Effects of ergocalciferol supplementation on the concentration of vitamin D and its metabolites in human milk. Journal of Nutrition, 119, 1639– 1646.
Tangpricha V, 2012. Vitamin D in food and supplements. American Journal of Clinical Nutrition, 95, 1299– 1300.
Taylor S, Wagner C and Hollis B, 2013. Lactation and vitamin D. In: RR Watson (ed). Handbook of Vitamin D in Human Health. Wageningen Academic Publishers, Wageningen, The Netherlands. pp. 632– 649.
Teotia M and Teotia SP, 1997. Nutritional and metabolic rickets. Indian Journal of Pediatrics, 64, 153– 157.
Thacher TD, 1997. Rickets without vitamin D deficiency in Nigerian children. Ambulatory Child Health, 3, 56– 64.
Thacher TD, Fischer PR, Pettifor JM, Lawson JO, Isichei CO and Chan GM, 2000. Case-control study of factors associated with nutritional rickets in Nigerian children. Journal of Pediatrics, 137, 367– 373.
Thacher TD, Obadofin MO, O'Brien KO and Abrams SA, 2009. The effect of vitamin D2 and vitamin D3 on intestinal calcium absorption in Nigerian children with rickets. Journal of Clinical Endocrinology and Metabolism, 94, 3314– 3321.
Thiebaud D, Burckhardt P, Costanza M, Sloutskis D, Gilliard D, Quinodoz F, Jacquet AF and Burnand B, 1997. Importance of albumin, 25(OH)-vitamin D and IGFBP-3 as risk factors in elderly women and men with hip fracture. Osteoporosis International, 7, 457– 462.
Thiele DK, Senti JL and Anderson CM, 2013. Maternal vitamin D supplementation to meet the needs of the breastfed infant: a systematic review. Journal of Human Lactation, 29, 163– 170.
Thiering E, Bruske I, Kratzsch J, Hofbauer LC, Berdel D, von Berg A, Lehmann I, Hoffmann B, Bauer CP, Koletzko S and Heinrich J, 2015. Associations between serum 25-hydroxyvitamin D and bone turnover markers in a population based sample of German children. Scientific Reports, 5, 18138.
Thompson SG and Higgins JP, 2002. How should meta-regression analyses be undertaken and interpreted? Statistics in Medicine, 21, 1559– 1573.
Thompson SG and Sharp SJ, 1999. Explaining heterogeneity in meta-analysis: a comparison of methods. Statistics in Medicine, 18, 2693– 2708.
Thompson GR, Lewis B and Booth CC, 1966. Absorption of vitamin D3-3H in control subjects and patients with intestinal malabsorption. Journal of Clinical Investigation, 45, 94– 102.
Toss G, Larsson L and Lindgren S, 1983. Serum levels of 25-hydroxyvitamin D in adults and elderly humans after a prophylactic dose of vitamin D2. Scandinavian Journal of Clinical and Laboratory Investigation, 43, 329– 332.
Trautvetter U, Neef N, Leiterer M, Kiehntopf M, Kratzsch J and Jahreis G, 2014. Effect of calcium phosphate and vitamin D3 supplementation on bone remodelling and metabolism of calcium, phosphorus, magnesium and iron. Nutrition Journal, 13, 6.
Tripkovic L, Lambert H, Hart K, Smith CP, Bucca G, Penson S, Chope G, Hyppönen E, Berry J, Vieth R and Lanham-New S, 2012. Comparison of vitamin D2 and vitamin D3 supplementation in raising serum 25-hydroxyvitamin D status: a systematic review and meta-analysis. American Journal of Clinical Nutrition, 95, 1357– 1364.
Trivedi DP, Doll R and Khaw KT, 2003. Effect of four monthly oral vitamin D3 (cholecalciferol) supplementation on fractures and mortality in men and women living in the community: randomised double blind controlled trial. British Medical Journal, 326, 469.
Tsiaras WG and Weinstock MA, 2011. Factors influencing vitamin D status. Acta Dermato-Venereologica, 91, 115– 124.
Uusi-Rasi K, Patil R, Karinkanta S, Kannus P, Tokola K, Lamberg-Allardt C and Sievanen H, 2015. Exercise and vitamin D in fall prevention among older women: a randomized clinical trial. JAMA Internal Medicine, 175, 703– 711.
Van Der Klis FR, Jonxis JH, Van Doormaal JJ, Sikkens P, Saleh AE and Muskiet FA, 1996. Changes in vitamin-D metabolites and parathyroid hormone in plasma following cholecalciferol administration to pre- and postmenopausal women in the Netherlands in early spring and to postmenopausal women in Curacao. British Journal of Nutrition, 75, 637– 646.
Vanlint S, 2013. Vitamin D and obesity. Nutrients, 5, 949– 956.
Vervel C, Zeghoud F, Boutignon H, Tjani JC, Walrant-Debray O and Garabedian M, 1997. Fortified milk and supplements of oral vitamin D. Comparison of the effect of two doses of vitamin D (500 and 1,000 UI/d) during the first trimester of life. Archives de Pédiatrie, 4, 126– 132.
Vestergaard P, Mosekilde L and Langdahl B, 2011. Fracture prevention in postmenopausal women. BMJ Clinical Evidence, 5, 1109.
Vieth R, 1999. Vitamin D supplementation, 25-hydroxyvitamin D concentrations, and safety. American Journal of Clinical Nutrition, 69, 842– 856.
Vieth R, Chan PCR and MacFarlane GD, 2001. Efficacy and safety of vitamin D3 intake exceeding the lowest observed adverse effect level. American Journal of Clinical Nutrition, 73, 288– 294.
Viljakainen HT, Natri AM, Karkkainen M, Huttunen MM, Palssa A, Jakobsen J, Cashman KD, Molgaard C and Lamberg-Allardt C, 2006a. A positive dose-response effect of vitamin D supplementation on site-specific bone mineral augmentation in adolescent girls: a double-blinded randomized placebo-controlled 1-year intervention. Journal of Bone and Mineral Research, 21, 836– 844.
Viljakainen HT, Palssa A, Karkkainen M, Jakobsen J and Lamberg-Allardt C, 2006b. How much vitamin D3 do the elderly need? Journal of the American College of Nutrition, 25, 429– 435.
Viljakainen HT, Vaisanen M, Kemi V, Rikkonen T, Kroger H, Laitinen EK, Rita H and Lamberg-Allardt C, 2009. Wintertime vitamin D supplementation inhibits seasonal variation of calcitropic hormones and maintains bone turnover in healthy men. Journal of Bone and Mineral Research, 24, 346– 352.
Viljakainen HT, Saarnio E, Hytinantti T, Miettinen M, Surcel H, Makitie O, Andersson S, Laitinen K and Lamberg-Allardt C, 2010. Maternal vitamin D status determines bone variables in the newborn. Journal of Clinical Endocrinology and Metabolism, 95, 1749– 1757.
Viljakainen HT, Korhonen T, Hytinantti T, Laitinen EK, Andersson S, Makitie O and Lamberg-Allardt C, 2011. Maternal vitamin D status affects bone growth in early childhood–a prospective cohort study. Osteoporosis International, 22, 883– 891.
Villareal DT, Civitelli R, Chines A and Avioli LV, 1991. Subclinical vitamin D deficiency in postmenopausal women with low vertebral bone mass. Journal of Clinical Endocrinology and Metabolism, 72, 628– 634.
Wabitsch M, Koletzko B and Moß A, 2011. Vitamin-D-Versorgung im Säuglings-, Kindes- und Jugendalter. Kurzfassung der Stellungnahme der Ernährungskommission der Deutschen Gesellschaft für Kinder- und Jugendmedizin (DGKJ) in Zusammenarbeit mit der Arbeitsgemeinschaft Pädiatrische Endokrinologie (APE). Monatsschrift Kinderheilkunde, 159, 766– 774.
Wagner CL, Hulsey TC, Fanning D, Ebeling M and Hollis BW, 2006. High-dose vitamin D3 supplementation in a cohort of breastfeeding mothers and their infants: a 6-month follow-up pilot study. Breastfeeding Medicine, 1, 59– 70.
Wagner CL, Taylor SN and Hollis BW, 2008a. Does vitamin D make the world go 'round'? Breastfeeding Medicine, 3, 239– 250.
Wagner CL, Greer FR, Academy American, American Academy of Pediatrics Section on Breastfeeding and American Academy of Pediatrics Committee on Nutrition, 2008b. Prevention of rickets and vitamin D deficiency in infants, children, and adolescents. Pediatrics, 122, 1142– 1152.
Wagner CL, Howard C, Hulsey TC, Lawrence RA, Taylor SN, Will H, Ebeling M, Hutson J and Hollis BW, 2010. Circulating 25-hydroxyvitamin D levels in fully breastfed infants on oral vitamin d supplementation. International Journal of Endocrinology, 2010, 235035.
Wagner D, Hanwell HE, Schnabl K, Yazdanpanah M, Kimball S, Fu L, Sidhom G, Rousseau D, Cole DE and Vieth R, 2011. The ratio of serum 24,25-dihydroxyvitamin D(3) to 25-hydroxyvitamin D(3) is predictive of 25-hydroxyvitamin D(3) response to vitamin D(3) supplementation. Journal of Steroid Biochemistry and Molecular Biology, 126, 72– 77.
Wagner CL, McNeil R, Hamilton SA, Winkler J, Rodriguez Cook C, Warner G, Bivens B, Davis DJ, Smith PG, Murphy M, Shary JR and Hollis BW, 2013a. A randomized trial of vitamin D supplementation in 2 community health center networks in South Carolina. American Journal of Obstetrics and Gynecology, 208, 137 e 1-13.
Wagner CL, McNeil RB, Johnson DD, Hulsey TC, Ebeling M, Robinson C, Hamilton SA and Hollis BW, 2013b. Health characteristics and outcomes of two randomized vitamin D supplementation trials during pregnancy: a combined analysis. Journal of Steroid Biochemistry and Molecular Biology, 136, 313– 320.
Wall CR, Stewart AW, Camargo CA Jr, Scragg R, Mitchell EA, Ekeroma A, Crane J, Milne T, Rowden J, Horst R and Grant CC, 2016. Vitamin D activity of breast milk in women randomly assigned to vitamin D3 supplementation during pregnancy. American Journal of Clinical Nutrition, 103, 382– 388.
Wallace AM, Gibson S, de la Hunty A, Lamberg-Allardt C and Ashwell M, 2010. Measurement of 25-hydroxyvitamin D in the clinical laboratory: current procedures, performance characteristics and limitations. Steroids, 75, 477– 488.
Wan X, Wang W, Liu J and Tong T, 2014. Estimating the sample mean and standard deviation from the sample size, median, range and/or interquartile range. BMC Medical Research Methodology, 14, 135.
Wang TJ, Zhang F, Richards JB, Kestenbaum B, van Meurs JB, Berry D, Kiel DP, Streeten EA, Ohlsson C, Koller DL, Peltonen L, Cooper JD, O'Reilly PF, Houston DK, Glazer NL, Vandenput L, Peacock M, Shi J, Rivadeneira F, McCarthy MI, Anneli P, de Boer IH, Mangino M, Kato B, Smyth DJ, Booth SL, Jacques PF, Burke GL, Goodarzi M, Cheung CL, Wolf M, Rice K, Goltzman D, Hidiroglou N, Ladouceur M, Wareham NJ, Hocking LJ, Hart D, Arden NK, Cooper C, Malik S, Fraser WD, Hartikainen AL, Zhai G, Macdonald HM, Forouhi NG, Loos RJ, Reid DM, Hakim A, Dennison E, Liu Y, Power C, Stevens HE, Jaana L, Vasan RS, Soranzo N, Bojunga J, Psaty BM, Lorentzon M, Foroud T, Harris TB, Hofman A, Jansson JO, Cauley JA, Uitterlinden AG, Gibson Q, Jarvelin MR, Karasik D, Siscovick DS, Econs MJ, Kritchevsky SB, Florez JC, Todd JA, Dupuis J, Hypponen E and Spector TD, 2010. Common genetic determinants of vitamin D insufficiency: a genome-wide association study. Lancet, 376, 180– 188.
Ward KA, Das G, Roberts SA, Berry JL, Adams JE, Rawer R and Mughal MZ, 2010. A randomized, controlled trial of vitamin D supplementation upon musculoskeletal health in postmenarchal females. Journal of Clinical Endocrinology and Metabolism, 95, 4643– 4651.
Waterhouse M, Tran B, Armstrong BK, Baxter C, Ebeling PR, English DR, Gebski V, Hill C, Kimlin MG, Lucas RM, Venn A, Webb PM, Whiteman DC and Neale RE, 2014. Environmental, personal, and genetic determinants of response to vitamin D supplementation in older adults. Journal of Clinical Endocrinology and Metabolism, 99, E1332– E1340.
Watson PE and McDonald BW, 2010. The association of maternal diet and dietary supplement intake in pregnant New Zealand women with infant birthweight. European Journal of Clinical Nutrition, 64, 184– 193.
Weaver CM and Heaney RP, 2006. Calcium. In: ME Shils, M Shike, AC Ross, B Caballero, RJ Cousins (eds.). Modern Nutrition in Health and Disease. Lippincott Williams & Wilkins, Baltimore, MD, Philadelphia, PA, USA. pp. 194– 210.
Weaver CM, McCabe LD, McCabe GP, Braun M, Martin BR, Dimeglio LA and Peacock M, 2008. Vitamin D status and calcium metabolism in adolescent black and white girls on a range of controlled calcium intakes. Journal of Clinical Endocrinology and Metabolism, 93, 3907– 3914.
Webb AR, 2006. Who, what, where and when-influences on cutaneous vitamin D synthesis. Progress in Biophysics and Molecular Biology, 92, 17– 25.
Webb AR and Engelsen O, 2006. Calculated ultraviolet exposure levels for a healthy vitamin D status. Photochemistry and Photobiology, 82, 1697– 1703.
Webb AR, DeCosta BR and Holick MF, 1989. Sunlight regulates the cutaneous production of vitamin D3 by causing its photodegradation. Journal of Clinical Endocrinology and Metabolism, 68, 882– 887.
Wei SQ, Audibert F, Hidiroglou N, Sarafin K, Julien P, Wu Y, Luo ZC and Fraser WD, 2012. Longitudinal vitamin D status in pregnancy and the risk of pre-eclampsia. BJOG: An International Journal of Obstetrics and Gynaecology, 119, 832– 839.
Wei SQ, Audibert F, Luo ZC, Nuyt AM, Masse B, Julien P, Fraser WD and Group MS, 2013. Maternal plasma 25-hydroxyvitamin D levels, angiogenic factors, and preeclampsia. American Journal of Obstetrics and Gynecology, 208, 390 e391-396.
Wharton B and Bishop N, 2003. Rickets. Lancet, 362, 1389– 1400.
White KM, Bauer SJ, Hartz KK and Baldridge M, 2009. Changes in body composition with yogurt consumption during resistance training in women. International Journal of Sport Nutrition and Exercise Metabolism, 19, 18– 33.
Whiting SJ, Calvo MS and Stephensen CB, 2013. Current understanding of vitamin D metabolism, nutritional status, and role in disease prevention. In: AM Coulston, CJ Boushey and MG Ferruzzi (eds.). Nutrition in the Prevention and Treatment of Disease. Academic Press, Elsevier, Burlington, MA, San Diego, CA, USA and London, UK. pp. 811– 838.
WHO/FAO (World Health Organization/Food and Agriculture Organization of the United Nations), 2004. Vitamin and mineral requirements in human nutrition: report of a joint FAO/WHO expert consultation, Bangkok, Thailand, 21–30 September 1998, 341 pp.
Wicherts IS, van Schoor NM, Boeke AJ, Visser M, Deeg DJ, Smit J, Knol DL and Lips P, 2007. Vitamin D status predicts physical performance and its decline in older persons. Journal of Clinical Endocrinology and Metabolism, 92, 2058– 2065.
Winzenberg T, Powell S, Shaw KA and Jones G, 2011. Effects of vitamin D supplementation on bone density in healthy children: systematic review and meta-analysis. British Medical Journal, 342, c7254.
Woo J, Lau E, Swaminathan R, Pang CP and MacDonald D, 1990. Biochemical predictors for osteoporotic fractures in elderly Chinese–a longitudinal study. Gerontology, 36, 55– 58.
Wood AD, Secombes KR, Thies F, Aucott LS, Black AJ, Reid DM, Mavroeidi A, Simpson WG, Fraser WD and Macdonald HM, 2014. A parallel group double-blind RCT of vitamin D3 assessing physical function: is the biochemical response to treatment affected by overweight and obesity? Osteoporosis International, 25, 305– 315.
Woodham PC, Brittain JE, Baker AM, Long DL, Haeri S, Camargo CA Jr, Boggess KA and Stuebe AM, 2011. Midgestation maternal serum 25-hydroxyvitamin D level and soluble fms-like tyrosine kinase 1/placental growth factor ratio as predictors of severe preeclampsia. Hypertension, 58, 1120– 1125.
Wortsman J, Matsuoka LY, Chen TC, Lu Z and Holick MF, 2000. Decreased bioavailability of vitamin D in obesity. American Journal of Clinical Nutrition, 72, 690– 693.
Yan L, Zhou B, Wang X, D'Ath S, Laidlaw A, Laskey MA and Prentice A, 2003. Older people in China and the United Kingdom differ in the relationships among parathyroid hormone, vitamin D, and bone mineral status. Bone, 33, 620– 627.
Young BE, McNanley TJ, Cooper EM, McIntyre AW, Witter F, Harris ZL and O'Brien KO, 2012. Maternal vitamin D status and calcium intake interact to affect fetal skeletal growth in utero in pregnant adolescents. American Journal of Clinical Nutrition, 95, 1103– 1112.
Yousefzadeh P, Shapses SA and Wang X, 2014. Vitamin D binding protein impact on 25-Hydroxyvitamin D levels under different physiologic and pathologic conditions. International Journal of Endocrinology, 2014, 981581.
Zeghoud F, Delaveyne R, Rehel P, Chalas J, Garabedian M and Odievre M, 1995. Vitamin D and pubertal maturation. Value and tolerance of vitamin D supplementation during the winter season. Archives de Pédiatrie, 2, 221– 226.
Zerwekh JE, 2008. Blood biomarkers of vitamin D status. American Journal of Clinical Nutrition, 87, 1087S– 1091S.
Zhang JY, Lucey AJ, Horgan R, Kenny LC and Kiely M, 2014. Impact of pregnancy on vitamin D status: a longitudinal study. British Journal of Nutrition, 112, 1081– 1087.
Zhou W, Langsetmo L, Berger C, Poliquin S, Kreiger N, Barr SI, Kaiser SM, Josse RG, Prior JC, Towheed TE, Anastassiades T, Davison KS, Kovacs CS, Hanley DA, Papadimitropoulos EA, Goltzman D and CaMos Research G, 2013. Longitudinal changes in calcium and vitamin D intakes and relationship to bone mineral density in a prospective population-based study: the Canadian Multicentre Osteoporosis Study (CaMos). Journal of Musculoskeletal and Neuronal Interactions, 13, 470– 479.
Zhu K, Devine A, Dick IM, Wilson SG and Prince RL, 2008a. Effects of calcium and vitamin D supplementation on hip bone mineral density and calcium-related analytes in elderly ambulatory Australian women: a five-year randomized controlled trial. Journal of Clinical Endocrinology and Metabolism, 93, 743– 749.
Zhu K, Bruce D, Austin N, Devine A, Ebeling PR and Prince RL, 2008b. Randomized controlled trial of the effects of calcium with or without vitamin D on bone structure and bone-related chemistry in elderly women with vitamin D insufficiency. Journal of Bone and Mineral Research, 23, 1343– 1348.
Zhu K, Austin N, Devine A, Bruce D and Prince RL, 2010. A randomized controlled trial of the effects of vitamin D on muscle strength and mobility in older women with vitamin D insufficiency. Journal of the American Geriatrics Society, 58, 2063– 2068.
Zittermann A and Koerfer R, 2008. Protective and toxic effects of vitamin D on vascular calcification: clinical implications. Molecular Aspects of Medicine, 29, 423– 432.

Abbreviations

1,25(OH)2D: 1,25-dihydroxy-vitamin D
1,25(OH)2D2: 1,25-dihydroxy-ergocalciferol
1,25(OH)2D3: 1,25-dihydroxy-cholecalciferol
1,24,25(OH)₃D: 1,24,25-trihydroxyvitamin D
25(OH)D: 25-hydroxy-vitamin D (sum of 25-hydroxy-vitamin D2 and 25-hydroxy-vitamin D3)
7-DHC: 7-dehydrocholesterol
aBMD: areal bone mineral density
ADL: activities of daily living
Afssa: Agence française de sécurité sanitaire des aliments
AHRQ: Agency for Healthcare Research and Quality
AI: adequate intake
ALP: alkaline phosphatase
AMD: age-related macular degeneration
AR: average requirement
BA: bone area
BioE: bioavailable estradiol
BioT: bioavailable testosterone
BMC: bone mineral content
BMD: bone mineral density
BMI: body mass index
BV: bone volume
CI: confidence interval
CPBA: competitive protein binding assay
CSA: cross-sectional area
CVD: cardiovascular disease
CYP: cytochrome P450
CYP24A1: 24-hydroxylase
CYP27B1: 1α-hydroxylase
CYP2R1, CYP27A1, CYP3A4, CYP2J3: 25-hydroxylase
D-A-CH: Deutschland-Austria-Confoederatio Helvetica
DBP: Vitamin D-binding protein
DEQAS: Vitamin D External Quality Assessment Scheme
DHCR7: 7-dehydrocholesterol reductase
DH: UK Department of Health
DRV: dietary reference value
DXA: dual-energy X-ray absorptiometry
EAR: estimated average requirement
ELISA: enzyme-linked immunosorbent assay
EOSP: early onset severe pre-eclampsia
ESPGHAN: European Society for Paediatric Gastroenterology, Hepatology and Nutrition
FAO: Food and Agriculture Organisation
FGF-23: fibroblast growth factor 23
GC: group specific component gene
HPLC: high-performance liquid chromatography
HR: hazard ratio
HR_adj: HR adjusted
I²: heterogeneity index
IOM: US Institute of Medicine of the National Academy of Sciences
IQR: interquartile range
ITT: intention-to-treat
IU: International unit
LC–MS: liquid chromatography–mass spectroscopy
LC–MS/MS: liquid chromatography-tandem mass spectroscopy
LMQ: leg muscle quality
NCM: Nordic Council of Ministers
NIST: National Institute of Standards and Technology
NNR: Nordic Nutrition Recommendations
NOAEL: No Observed Adverse Effect Level
OR: odds ratio
OV: osteoid volume
PI: prediction interval
pQCT: peripheral quantitative computed tomography
PRI: population reference intake
PTH: parathyroid hormone
Q1: first quartile
QCT: quantitative computed tomography
QUS: quantitative ultrasound
RCT: randomised controlled trial
RDA: recommended dietary allowance
RI: Recommended Intake
RIA: radioimmunoassay
RMP: Reference measurement procedure
RNI: Reference Nutrient Intake
RoB: risk of bias
RR: relative risk
SACN: Scientific Advisory Committee on Nutrition
SCF: Scientific Committee for Food
SD: standard deviation
SGA: small-for-gestational-age
SH: sex hormones
SHBG: sex hormone binding globulin
SNP: single nucleotide polymorphism
SPPB: short physical performance battery
TUAG: Timed Up And Go
UHT: ultra-high temperature
UL: tolerable upper intake level
UV: ultraviolet
VDR: Vitamin D receptor
VDSP: Vitamin D standardization program
Vitamin D₂: ergocalciferol
Vitamin D₃: cholecalciferol
WHO: World Health Organization

Appendix A – Measurements for the assessment of bone healt

Bone measurements in children and adults may be obtained using different techniques of bone densitometry, e.g. dual-energy X-ray absorptiometry (DXA), quantitative computed tomography (QCT), peripheral quantitative computed tomography (pQCT) or quantitative ultrasound (QUS). Assessments of the advantages, precision, specificity and sensitivity of these methods in different populations (for example: Baroncelli, 2008; Brunner et al., 2011; Edelmann-Schafer et al., 2011) and recommendations on their use (e.g. from the International Society for Clinical Densitometry) have been published.

DXA is the most commonly used method of measuring bone mass. DXA measurements may include lumbar spine, hip, forearm and whole body. The DXA scans provide a number of outcomes: bone area, bone mineral content (BMC) and bone mineral density (BMD) in the above-mentioned anatomical areas. BMD is a two-dimensional measurement of the bone, i.e. areal BMD (aBMD, g/cm²). The calibration of the different DXA densitometers may differ between studies, resulting in different BMD and BMC values.

In contrast, QCT, which also involves X-ray radiation, is used to measure three-dimensional (volumetric) BMD (g/cm³) in the spine or hip, and to assess bone structure, i.e. separately analyse BMD for the compact (or cortical) bone or for the trabecular (or cancellous) bone. Moreover, pQCT measures bone characteristics in 'peripheral' body sites, such as the forearms or legs, and provides a number of outcomes, e.g. volumetric BMD, the stress–strain index and measures of the geometry of the bone (i.e. spatial distribution of the bone mass) (Section 5.1.2). QUS methods have been developed to give estimates of bone health, without the use of ionising radiation. Measurements are usually performed at the heel (calcaneus). In its review, the Panel did not identify any recent relevant study on bone-related outcomes using this technique.

Appendix B – Summary of the evidence considered by the IOM to set DRVs for vitamin D

1 Adults

IOM (2011) used mostly the systematic reviews by Cranney et al. (2007) and by Chung et al. (2009) to draw conclusions on 25(OH)D concentrations and bone-related health outcomes.

Cranney et al. (2007) considered 19 studies on the association between serum 25(OH)D concentrations and BMD in older adults. They comprised six randomised controlled trial (RCTs) on vitamin D supplementation with calcium (Dawson-Hughes et al., 1995; Storm et al., 1998; Schaafsma et al., 2002; Cooper et al., 2003; Aloia et al., 2005) or without calcium (Ooms et al., 1995). These RCTs and two cohort studies (Dennison et al., 1999; Gerdhem et al., 2005) reported no significant association between serum 25(OH)D concentrations and BMD or bone loss. However, five other cohort studies reported a significant association, particularly at the hip sites (Rosen et al., 1994; Stone et al., 1998; Melin et al., 2001; del Puente et al., 2002; Bischoff-Ferrari et al., 2005), and only one at the lumbar spine (Rosen et al., 1994). Six case–control studies (Villareal et al., 1991; Thiebaud et al., 1997; Boonen et al., 1999; Landin-Wilhelmsen et al., 1999; Yan et al., 2003; Al-oanzi et al., 2006) reported an association between 25(OH)D concentrations and BMD, most consistently at the femoral neck. Chung et al. (2009) included two additional RCTs (Andersen et al., 2008, Zhu et al., 2008b). Zhu et al. (2008b) showed that vitamin D₂ supplementation over 1 year provided no extra benefit in older Caucasian women (mean baseline serum 25(OH)D concentration: 44.3 nmol/L) on total hip BMD compared to calcium supplementation alone. Andersen et al. (2008) reported no effect of the vitamin D₃ supplementation on BMC/BMD and no differences in 1-year BMD changes at the lumbar spine between the intervention and placebo groups, either in female or in male Pakistani immigrants in Denmark (mean baseline serum 25(OH)D concentration: 12 (women) and 21 (men) nmol/L).

With regard to vitamin D supplementation with or without calcium in older adults and BMD, Cranney et al. (2007) identified 17 RCTs (Dawson-Hughes et al., 1991, 1995, 1997; Chapuy et al., 1992, 2002; Ooms et al., 1995; Baeksgaard et al., 1998; Komulainen et al., 1998; Hunter et al., 2000; Patel et al., 2001; Jensen et al., 2002; Cooper et al., 2003; Grados et al., 2003; Harwood et al., 2004; Meier et al., 2004; Aloia et al., 2005; Jackson et al., 2006), mostly in post-menopausal women and older men (a few references (Patel et al., 2001; Meier et al., 2004) also included younger subjects). Combining results of individual studies to calculate weighted mean differences, Cranney et al. (2007) concluded that vitamin D₃ plus calcium supplementation compared with placebo resulted in 'small' significant increases in BMD of the lumbar spine, total body and femoral neck (but not of the forearm). However, they concluded that vitamin D₃ plus calcium compared with calcium did not have a significant effect on BMD of the lumbar spine, total hip, forearm or total body (but the effect for femoral neck was significant). They also concluded that vitamin D₃ supplementation alone versus placebo had a significant effect on BMD at the femoral neck but not at the forearm. Chung et al. (2009) identified three additional RCTs in older adults (Moschonis and Manios, 2006; Bolton-Smith et al., 2007; Zhu et al., 2008a), only two of which (Moschonis and Manios, 2006; Zhu et al., 2008a) found a significant increase in hip or total BMD in post-menopausal women receiving vitamin D₂ or D₃ plus calcium compared with placebo.

For osteomalacia, the IOM used a study on post-mortem biopsies (Priemel et al., 2010) (Section 5.1.2.1.2).

For fracture risk in older adults, with regard to serum 25(OH)D concentrations, Cranney et al. (2007) identified only observational studies. They took into account three prospective cohort studies in independently living older adults (Woo et al., 1990; Cummings et al., 1998; Gerdhem et al., 2005). They also considered case–control studies (Lund et al., 1975; Lips et al., 1983, 1987; Punnonen et al., 1986; Cooper et al., 1989; Lau et al., 1989; Boonen et al., 1997, 1999; Thiebaud et al., 1997; Diamond et al., 1998; Landin-Wilhelmsen et al., 1999; LeBoff et al., 1999; Erem et al., 2002; Bakhtiyarova et al., 2006). Cranney et al. (2007) concluded that there was inconsistent evidence for an association between a lower serum 25(OH)D concentration and an increased risk of fracture. IOM (2011) identified six additional observational studies (Cauley et al., 2008, 2010; Looker and Mussolino, 2008; van Schoor et al., 2008; Ensrud et al., 2009; Melhus et al., 2010). These showed inconsistent results on 25(OH)D concentrations below which there may be an increased risk of fracture, which varied between 30 and 70 nmol/L.

With regard to vitamin D supplementation and risk of fractures, Cranney et al. (2007) assessed 15 RCTs (Chapuy et al., 1992, 2002; Lips et al., 1996; Dawson-Hughes et al., 1997; Komulainen et al., 1998; Pfeifer et al., 2000; Trivedi et al., 2003; Anderson et al., 2004; Harwood et al., 2004; Larsen et al., 2004; Flicker et al., 2005; Grant et al., 2005; Porthouse et al., 2005; Jackson et al., 2006; Law et al., 2006). These RCTs investigated the effect of vitamin D (with or without calcium) on fractures in post-menopausal women and older men with baseline 25(OH)D concentrations ranging from 22 to 82.7 nmol/L. Eleven of these RCTs used vitamin D₃ preparations (7.5–20 μg/day), and the others vitamin D₂ (Anderson et al., 2004; Larsen et al., 2004; Flicker et al., 2005; Law et al., 2006). Cranney et al. (2007) conducted a meta-analysis of 13 of these RCTs, omitting the abstract by Anderson et al. (2004) and the study by Larsen et al. (2004) with no placebo control. Cranney et al. (2007) calculated combined ORs that indicated non-significant effect of the interventions for total fractures,33 non-vertebral fractures,34 hip fractures,35 vertebral fractures,36 and total or hip fractures in community-dwelling older adults. Combined ORs also indicated significant reduction in the risk of fractures for end of study 25(OH)D concentration ≥ 74 nmol/L (compared to 25(OH)D < 74 nmol/L),37 and for total or hip fractures in institutionalised older adults.38 Chung et al. (2009) identified three additional RCTs on bone health (Bunout et al., 2006; Burleigh et al., 2007; Lyons et al., 2007), two of which investigated fracture risk. These did not show significant effects of either vitamin D₂ (four-monthly dose equivalent to 20.6 μg/day) compared with placebo, or of vitamin D₃ (20 μg/day) plus calcium compared with calcium, in reducing the risk of total fractures, in a cohort of hospital inpatients (Burleigh et al., 2007) and in older adults living in residential or care homes (Lyons et al., 2007). IOM (2011) identified two additional RCTs (Salovaara et al., 2010; Sanders et al., 2010). In both studies, there was no statistically significant effect of the combination of calcium and vitamin D₃ on incident fractures compared to no treatment.

Based on Cranney et al. (2007) and Chung et al. (2009) and observational data outside of these reviews (four other cross-sectional (Bischoff-Ferrari et al., 2004; Boxer et al., 2008; Stewart, 2009) or longitudinal (Wicherts et al., 2007) observational studies), IOM (2011) found that there was some support for an association between 25(OH)D concentrations and physical performance (data for this outcome were considered together with that for the risk of falls mentioned below). However, IOM (2011) found that high-quality and large observational cohort studies were lacking, and that randomised trials suggest that vitamin D dosages of at least 20 μg/day, with or without calcium, may improve physical performance measures. Although the IOM considered that high doses of vitamin D (i.e. ≥ 20 μg/day) may provide greater benefit for physical performance than low doses (i.e. 10 μg/day), the IOM found that the evidence was insufficient to define the shape of the dose–response curve for higher levels of intake.

Based on Cranney et al. (2007) and Chung et al. (2009) and two RCTs (Bischoff-Ferrari et al., 2010; Sanders et al., 2010) published afterwards, IOM (2011) considered that no consistent result was found from randomised trials that tested for effects of vitamin D with and without calcium on reduction in risk for falls. IOM considered 20 randomised trials on oral doses (Graafmans et al., 1996; Pfeifer et al., 2000, 2009; Chapuy et al., 2002; Bischoff-Ferrari et al., 2003, 2006, 2010; Trivedi et al., 2003; Flicker et al., 2005; Grant et al., 2005; Larsen et al., 2005; Law et al., 2006; Broe et al., 2007; Burleigh et al., 2007; Prince et al., 2008) or injected doses (Latham et al., 2003; Dhesi et al., 2004; Harwood et al., 2004; Smith et al., 2007; Sanders et al., 2010). These RCTs had heterogeneous designs, e.g. subjects were either free-living or institutionalised older subjects, and supplemented with vitamin D with or without calcium and compared to calcium or placebo. From these, IOM noted that only four (Pfeifer et al., 2000; Harwood et al., 2004; Flicker et al., 2005; Broe et al., 2007) found a significant effect of vitamin D on fall incidence, and that there were only two significant studies for fallers (Pfeifer et al., 2000, 2009).39 The IOM (2011) noted that a number of the RCTs analysed falls rather than fallers. The IOM concluded that the greater part of the causal evidence indicated no significant reduction in fall risk related to vitamin D intake or achieved concentration in blood. IOM (2011) noted that Cranney et al. (2007)40 and Chung et al. (2009) found no consistency between study findings. With regard to the evidence from observational studies, the IOM noted one longitudinal Dutch study (Snijder et al., 2006) (which was not part of Cranney et al. (2007) or Chung et al. (2009)) that found that a serum 25(OH)D concentration < 25 nmol/L was independently associated with an increased risk of falling for subjects who experienced two or more falls compared with those who did not fall or fell once. IOM (2011) summarised that observational studies suggested an association between a higher serum 25(OH)D concentration and a lower risk of falls in older adults.

In relation to calcium absorption in adults, IOM (2011) considered RCTs in mainly post-menopausal women with vitamin D supplementation (Francis et al., 1996; Patel et al., 2001; Zhu et al., 2008a,b), using the dual isotope technique. The RCTs varied considerably in design and, overall, showed no effect of increasing the serum 25(OH)D concentrations on intestinal calcium absorption compared with placebo. In a short-term RCT in post-menopausal women using the dual isotope technique, Hansen et al. (2008) showed a 3% increase in absorption after raising the serum 25(OH)D concentration from 55 to 160 nmol/L. IOM also considered cross-sectional studies using the single isotope technique (Kinyamu et al., 1998; Devine et al., 2002; Heaney et al., 2003a; Need et al., 2008; Aloia et al., 2010). In particular, in 319 patients (mostly men) attending osteoporosis clinics and with serum 25(OH)D concentrations less than 40 nmol/L, Need et al. (2008) found no increase in fractional calcium absorption in subjects with serum 25(OH)D concentrations above 10 nmol/L. The studies by Heaney et al. (2003a) and Kinyamu et al. (1998) indicated no changes in fractional calcium absorption across ranges of serum 25(OH)D concentrations of 60–154 nmol/L and 50–116 nmol/L, respectively. In the study by Aloia et al. (2010) in 492 African American and 262 Caucasian women (20–80 years), no relationship was found between calcium absorption and serum 25(OH)D concentrations ranging from 30 to 150 nmol/L. The relationship between calcium absorption and 1,25(OH)₂D concentration was positive and stronger for lower than for higher 25(OH)D concentrations.

IOM (2011) concluded that serum 25(OH)D concentrations of 40 nmol/L, 50 nmol L or higher were sufficient to meet bone health requirements for most adults in RCTs, and to provide maximal population coverage in observational studies on adults and bone health.

2 Infants and children

For infants, Cranney et al. (2007) reported on the inconsistent results of two RCTs with vitamin D₂ supplementation examining serum 25(OH)D concentrations and BMC (Greer et al., 1982; Greer and Marshall, 1989), and on the inconsistent results of three case–control studies (Bougle et al., 1998; Namgung et al., 1998; Park et al., 1998) examining serum 25(OH)D concentrations and BMD and/or BMC. Chung et al. (2009) found no additional RCTs in infants.

For children, Cranney et al. (2007) identified three RCTs (Ala-Houhala et al., 1988b; El-Hajj Fuleihan et al., 2006; Viljakainen et al., 2006a), two prospective cohort studies (Lehtonen-Veromaa et al., 2002; Javaid et al., 2006), and one case–control study (Marwaha et al., 2005). In children (8–10 years) receiving vitamin D₂ supplementation or placebo for more than 1 year (Ala-Houhala et al., 1988b), the change in serum 25(OH)D concentrations after supplementation was not accompanied by a change in distal radial BMC. However, Cranney et al. (2007) reported, in girls (10–17 years) receiving two doses of vitamin D₃ supplementation or a placebo for 1 year (El-Hajj Fuleihan et al., 2006), that baseline serum 25(OH)D concentrations were significantly related to baseline BMD (positively) or per cent change in BMC (negatively), at the lumbar spine, femoral neck, and radius. They also reported for this study a significant increase in BMC only of the total hip in girls receiving the highest dose of supplementation, compared with placebo (El-Hajj Fuleihan et al., 2006). In girls (11–12 years) with 'adequate' calcium intake and who received one of two doses of daily vitamin D₃ supplementation or a placebo for 1 year, mean achieved serum 25(OH)D was above 50 nmol/L in both intervention groups (Viljakainen et al., 2006a). A significant increase in BMC of the femur (for both doses) or lumbar spine (for the highest dose) was reported in subjects with compliance above 80 %, but this was not statistically significant in the ITT analysis. Cranney et al. (2007) reported a positive association between baseline serum 25(OH)D concentrations of girls (9–15 years) followed for 3 years and change in BMD (Lehtonen-Veromaa et al., 2002), and between maternal serum 25(OH)D during pregnancy and BMC of the children (8–9 years) (Javaid et al., 2006). However, there was no significant correlation between serum 25(OH)D and BMD of children (10–18 years) in either group of the case–control study (Marwaha et al., 2005).

Cranney et al. (2007) concluded that there was evidence of an association between serum 25(OH)D concentrations and baseline BMD and change in BMD or related variables, but that the results of RCTs were not consistent with regard to the effect of vitamin D supplementation on BMD or BMC across skeletal sites and age groups. Chung et al. (2009) identified one RCT in 26 healthy Pakistani immigrant girls (10–17 years) living near Copenhagen (mean baseline 25(OH)D concentration: 11 nmol/L), and receiving one of two doses of vitamin D₃ supplementation alone or a placebo (Andersen et al., 2008). There were no significant differences in whole-body BMC changes between the supplemented groups and the placebo group. Chung et al. (2009) identified another RCT (Cheng et al., 2005) in healthy girls (10–12 years) (mean baseline 25(OH)D concentration: 35 nmol/L) receiving supplementation with vitamin D₃ and calcium or a placebo, which showed no significant difference in BMC changes between groups after 2 years.

According to IOM (2011) and Cranney et al. (2007), among 13 studies on rickets, six (including one RCT (Cesur et al., 2003)) reported mean or median serum 25(OH)D concentrations below 27.5 nmol/L, and expressed as about 30 nmol/L, in children with rickets (Garabedian et al., 1983; Markestad et al., 1984; Bhimma et al., 1995; Majid Molla et al., 2000; Cesur et al., 2003; Dawodu et al., 2005). The others (before-after or case–control) studies were reported as showing mean/median serum 25(OH)D concentrations higher than 30 nmol/L and up to 50 nmol/L in children with rickets (Arnaud et al., 1976; Elzouki et al., 1989; Oginni et al., 1996; Thacher, 1997; Thacher et al., 2000; Balasubramanian et al., 2003; Graff et al., 2004). Seven case–control studies showed lower serum 25(OH)D concentrations in cases than in controls (Arnaud et al., 1976; Oginni et al., 1996; Majid Molla et al., 2000; Thacher et al., 2000; Balasubramanian et al., 2003; Graff et al., 2004; Dawodu et al., 2005). Three studies were conducted in Western countries (Arnaud et al., 1976; Garabedian et al., 1983; Markestad et al., 1984), while most were conducted in non-Western countries with low calcium intake. Cranney et al. (2007) noted that low calcium intake can influence the relationship between serum 25(OH)D and rickets and that the 25(OH)D cut-off value for rickets in populations with high calcium intake is unclear. Chung et al. (2009) did not identify any additional study on rickets.

For children, IOM (2011) identified two dual isotope studies (an observational study (Abrams et al., 2009) or a randomised trial (Thacher et al., 2009)) on fractional calcium absorption, and a pooled analysis of several 3-week calcium-balance metabolic studies in 105 girls (11–15 years) (Weaver et al., 2008), in which serum 25(OH)D concentration was not related to net calcium absorption or retention. However, in this last study, calcium balance or retention was calculated by subtracting calcium excretion through urine and faeces from dietary calcium intake. Pooling studies in 251 children (about 5–17 years) and assessing the relationship of 25(OH)D concentration (as a continuous variable) with either fractional or total calcium absorption, according to pubertal status and/or calcium intake, Abrams et al. (2009) found inconsistent results. However, when 25(OH)D was studied as a categorical variable in the whole population, fractional calcium absorption adjusted (in particular) for calcium intake was slightly, but significantly (p < 0.05), higher at 25(OH)D concentration of 28–50 nmol/L, compared with ranges of 50–80 nmol/L or greater than 80 nmol/L. In Nigeria, 17 prepubertal children, with rickets, 'low' calcium intake and mean baseline 25(OH)D concentration of 50 nmol/L, were randomised to receive single oral supplementation of vitamin D₂ or D₃ (Thacher et al., 2009). An increase in serum 25(OH)D concentrations was reported in both groups, but at 'low' calcium intake and with no significant increase in fractional calcium absorption between baseline and 3 days after supplementation (Thacher et al., 2009).

3 Pregnancy

For IOM (2011), during pregnancy, maternal 1,25(OH)₂D increases, while 25(OH)D is generally unaffected in unsupplemented women. Animal data reviewed by IOM (2011) suggested that the increased calcium absorption during pregnancy is independent from vitamin D or 1,25(OH)₂D, and observational data showed that vitamin D-deficiency rickets may develop weeks or months after birth. For maternal bone health during pregnancy, Cranney et al. (2007) identified two prospective observational studies (Ardawi et al., 1997; Morley et al., 2006) and one before-and-after study (Datta et al., 2002), which found either a negative or no correlation between maternal serum 25(OH)D and parathyroid hormone (PTH) concentrations. Maternal BMD/BMC was not investigated in these studies. Chung et al. (2009) or IOM (2011) identified no RCTs for this outcome.

For the prevention of pre-eclampsia, the IOM noted the absence of placebo-controlled randomised trials in favour of an effect of vitamin D. One RCT (Marya et al., 1987) (identified by Chung et al. (2009)) found no effect of vitamin D and calcium supplementation on the incidence of pre-eclampsia and the results of a non-randomised trial on vitamin D₃ and calcium supplementation (Ito et al., 1994) were found unclear. Two observational studies showed inverse associations between vitamin D intake from supplements and risk of pre-eclampsia (Hypponen et al., 2007; Haugen et al., 2009). For the IOM, case–control or nested case–control studies (including one (Bodnar et al., 2007) found by Chung et al. (2009)), investigating serum 25(OH)D concentration and the risk of pre-eclampsia or comparing serum 25(OH)D concentration in women with or without pre-eclampsia, found contradictory results (Frolich et al., 1992; Seely et al., 1992; Bodnar et al., 2007). However, one case–control study (Lalau et al., 1993) showed lower total or free serum 1,25(OH)₂D in women with pregnancy-induced hypertension.

The IOM noted the limited observational evidence on non-skeletal maternal outcomes (caesarean section, obstructed labour, vaginosis), reviewed neither in Cranney et al. (2007) nor in Chung et al. (2009). In RCTs (most identified by Chung et al. (2009)) on maternal vitamin D supplementation and birth weight or length (Brooke et al., 1980; Maxwell et al., 1981; Mallet et al., 1986; Marya et al., 1988), no effect was observed. IOM also reported on observational studies with conflicting results on vitamin D intake/status during pregnancy and infant birth size or small-for-gestational age measurements (Brunvand et al., 1998; Morley et al., 2006; Gale et al., 2008; Farrant et al., 2009; Scholl and Chen, 2009; Bodnar et al., 2010; Leffelaar et al., 2010).

For fetal/newborn bone health, an RCT (Delvin et al., 1986) was reported as showing no effect of maternal vitamin D supplementation on fetal calcium homeostasis. The IOM also considered observational studies (Maxwell and Miles, 1925; Brooke et al., 1980; Congdon et al., 1983; Silver et al., 1985; Pereira and Zucker, 1986; Campbell and Fleischman, 1988; Specker et al., 1992; Specker, 1994; Takeda et al., 1997; Teotia and Teotia, 1997; Kitanaka et al., 1998; Akcakus et al., 2006; Bouillon et al., 2006; Beck-Nielsen et al., 2009). From them, the IOM concluded that there was no relationship between maternal 25(OH)D concentration and fetal BMC or BMD, as well as normal fetal skeletal development and no radiological evidence of rickets at birth in case of maternal vitamin D 'deficiency' or the absence of 1α-hydroxylase or the vitamin D receptor (VDR). Other observational studies were reported as showing lower maternal and neonatal serum 25(OH)D concentrations in infants with craniotabes (Reif et al., 1988) and an inverse association between fetal femur metaphyseal cross-sectional area or splaying index and maternal 25(OH)D during pregnancy (Mahon et al., 2010). From another observational study (Viljakainen et al., 2010), the IOM noted the lower newborn tibia BMC and cross-sectional area with maternal serum 25(OH)D concentration below 42.6 nmol/L (mean of first trimester and two-day post-partum values, close to the 'EAR-type value' proposed by the IOM), compared to higher serum 25(OH)D, after adjustments for potential confounders.

Regarding the relationship between maternal 25(OH)D during pregnancy and childhood bone health, the IOM refers to a study providing follow-up data on 33% of the children included in a mother-infant cohort (n = 596 initially) (Javaid et al., 2006). This observational study reported a positive association between whole-body and lumbar spine BMC and aBMD in children (9 years) and maternal serum 25(OH)D concentrations in pregnancy (mean: 34 weeks) after adjustments for potential confounders. Children of mothers whose serum 25(OH)D concentrations in pregnancy were less than 27.5 nmol/L (compared to above 50 nmol/L) had a significantly lower whole-body BMC (p = 0.002).

4 Lactation

IOM (2011) stated that breast milk is not a significant source of vitamin D for breastfed infants, and that the maternal skeleton recovers BMC after the end of lactation. IOM (2011) considered observational studies (Cancela et al., 1986; Okonofua et al., 1987; Kent et al., 1990; Alfaham et al., 1995; Cross et al., 1995; Sowers et al., 1998; Ghannam et al., 1999) and intervention studies (Greer et al., 1982; Rothberg et al., 1982; Ala-Houhala, 1985; Ala-Houhala et al., 1988b; Greer and Marshall, 1989; Takeuchi et al., 1989; Kalkwarf et al., 1996; Hollis and Wagner, 2004a; Basile et al., 2006; Wagner et al., 2006; Saadi et al., 2007). Some of these had been identified by Cranney et al. (2007) and Chung et al. (2009). From these studies, the IOM reported no major change in serum 25(OH)D concentration during lactation compared to non-lactating women, and that providing vitamin D to lactating mothers increased their serum 25(OH)D concentrations, without significant effect on either infant serum 25(OH)D concentrations (for supplementation below 100 μg/day) or infant weight or height. The IOM also noted the lack of association between maternal 25(OH)D concentration and maternal post partum changes in BMD (e.g. lumbar spine or femoral neck), or breast milk calcium content (Prentice et al., 1997). IOM (2011) noticed that no RCTs had investigated the influence of maternal vitamin D intake or status on the recovery of maternal skeletal mineral content after the end of lactation.

Appendix C – Dose–response analysis undertaken by EFSA of serum 25(OH)D to total vitamin D intake: methods and key results

The specific objective of the quantitative analysis was to estimate the dose–response relationship between vitamin D total intake and plasma/serum 25(OH)D concentration in situations of assumed minimal endogenous vitamin D synthesis through exposure to the sun or artificial ultraviolet (UV) irradiation in the healthy population.

The analysis as detailed in the present Appendix was developed based on the related analysis plan, which has been informed by the systematic review protocol drafted by the contractor (Brouwer-Brolsma et al., 2016) in agreement with EFSA and by specific input from the NDA WG on dietary reference values for Vitamins.

Data synthesis: meta-analyses, meta-regression, dose–response models

1 Criteria under which study data were quantitatively synthesised

In a meta-analytic approach, quantitative synthesis is usually carried out if included studies are sufficiently homogeneous to allow for meaningful combined estimates.

In the context of the current analysis, a high statistical heterogeneity across included studies was expected; the relative contributions of methodological heterogeneity and/or 'clinical' heterogeneity were evaluated by analysing the relevant data extracted at the study level (e.g. dimensions of methodological quality, intake-status influencing factors).

In recognition of such heterogeneity, prospective observational studies were analysed separately from randomised trials, the latter being the basis for the dose–response modelling.

Once the methodological heterogeneity possibly due to differences in the internal validity of the results from individual studies is characterised, the remaining variation is likely to reflect a real phenomenon that describes the extent to which different populations behave differently. Independently of the extent to which identified 'clinical' covariates could explain it, heterogeneity was incorporated in the derivation of DRVs, in the idea that they are being applied to different populations in different contexts.

The very high heterogeneity was taken into account in meta-analyses and meta-regressions applying a random-effects model. A random-effects model assumes that true effects follow a normal distribution around a pooled weighted mean (or around the conditional linear predictor for models) and allows for the residual heterogeneity among responses not characterised by subgroups analyses (or not modelled by the explanatory variables included in the multivariable models).

All statistical analyses were performed with STATA version 13.1 (Stata-Corp, College Station, TX, USA). Unless otherwise specified, all estimates were presented with 95% confidence intervals (CIs) and all analyses were carried out at the level of statistical significance of 0.05.

2 Summary measures

The continuous outcome (i.e. plasma/serum 25(OH)D as a marker of vitamin D status) was analysed using the summary data extracted by the contractor (Brouwer-Brolsma et al., 2016) for each arm in each individual study: the number of participants included (and assessed); the mean values and SDs of the baseline and final values of 25(OH)D (as reported in the original paper or as converted by the contractor to nmol/L) at each relevant time point (i.e. final concentrations measured in a period of assumed minimal endogenous vitamin D synthesis) and for each vitamin D dose/intake (up to 50 μg/day dose).

Summary measures and related standard errors were either calculated or imputed based on the type of summary data available (e.g. means were estimated from medians when these were available). Absolute achieved means and their standard errors were meta-analysed and used in the dose–response meta-regression models. Weighted mean differences (with 95% CI) as calculated by pooling study-specific estimates (when a control arm was available) in random-effects meta-analyses were used for comparative purposes. Net changes from baseline to achieved means by arm were calculated to check for consistency of results and to identify heterogeneity potentially due to methodological issues.

3 Unit of analysis issues

All included trials were assessed in order to check whether the unit of randomisation was consistent with the unit of analysis in the trial (i.e. per individual randomised).

Only one cross-over trial was initially included (Patel et al., 2001), which was treated according to the contractor's criteria (i.e. only the two periods from November through February were considered eligible and extracted as two different studies: Patel et al., 2001a and Patel et al., 2001b). The trial was subsequently excluded based on its design and net change values (Table 8 of Appendix D.A).

4 Dealing with missing data

The contractor contacted the original authors of the individual studies to obtain relevant missing data; imputation was used in the current analysis (e.g. mean age derived from age range) to deal with key summary information that could not be retrieved despite the contractor's efforts.

Specific formulae (Higgins and Green, 2011) were applied to derive summary data where not directly extracted/available in the format of the statistics mentioned in Section 2 of this Appendix (e.g. SDs were calculated from standard errors and group size or CIs). If no calculation/estimation was possible, the missing data were imputed according to the approach proposed by Wan et al. (2014).

Information for all relevant study-level characteristics was complete with the exceptions of funding source (6% missing), ethnicity (47%) and mean Body Mass Index (28%) (Appendix D.B, Tables 9–11). Availability of BMI mean values in the final data set was maximised by calculating it from mean weight and mean height (BMI = body weight (kg)/height² (meters)) when available; missing data proportion dropped to 16%. While developing the final model, BMI missing data were included in a specific category as 'not reported', to be able to compare models with and without BMI as covariate (i.e. assuring same number of arms in all models). Funding and ethnicity were analysed likewise, although the high proportion of missing values for ethnicity prevented it from being included in the final model.

Background intake estimates were added to the supplemental vitamin D dose to generate total vitamin D intake estimates. If the habitual vitamin D intake of the cohort(s) within a study was not reported, surrogates were imputed using the appropriate age- and sex- specific mean vitamin D intake values (from food) from the national nutrition survey relevant to the country in which the study was performed (17 studies – Appendix D.B, Table11) ; values were weighted for the arm-specific sex proportions and age ranges.

Only for one trial (Rich-Edwards et al., 2011) on children from Mongolia values were imputed from another included trial (Madsen et al., 2013) on children from Denmark, as participants were of comparable age.

Sensitivity analyses to assess the impact of summary data and background intake imputations on the overall analyses were performed; the intake coefficient estimated in the dose–response model with no covariates on the revised data did not change substantially from the intake coefficient on the original values, showing an overall minor impact of imputation on the crude dose–response relationship.

5 Assessment of heterogeneity

Statistical heterogeneity was tested using the χ² test (Cochran's Q test; significance level: 0.10) and quantified by calculating the I² statistic (Higgins and Thompson, 2002).

I ² ranges between 0 and 100 per cent and quantifies the proportion of the variability in effect estimates that can be attributed to heterogeneity rather than chance. As a reference, 0% to 40% might not be important; 30% to 60% may represent moderate heterogeneity; 50% to 90% may represent substantial heterogeneity; 75% to 100% represents considerable heterogeneity (Higgins and Green, 2011).

I² was 99% in the overall meta-analysis of achieved mean values and did not drop below 94% in any subgroups except when intervention doses were investigated (85% in trials with dose = 2 μg/day, 76% in trials with dose = 50 μg/day). Given the very high level of heterogeneity between trials, possible sources were explored by subgroup analysis, meta-regression and/or sensitivity analysis.

6 Data checking

For each variable, the proportion of missing observations was calculated and range checks carried out to ensure that all values were plausible. The distributions of continuous variables were explored graphically and the frequency distributions of categorical variables tabulated. Key variables were cross-tabulated or scattered against each other to check for consistency. Summary data were double checked against original publications whenever deemed necessary and unit conversions of all included 25(OH)D and vitamin D dose/intake values were verified (ng/mL converted to nmol/L by multiplying by 2.496; IU/day converted to μg/day by dividing by 40).

7 Meta-analyses

Random-effects meta-analyses of summary response measures were carried out using the DerSimonian and Laird approach (DerSimonian and Laird, 1986), which encompasses both variability due to chance (i.e. the within-study variance component in the denominator of the individual study weight) and variability due to heterogeneity (i.e. the between-study variance component added in the denominator of the individual study weight – T ² statistic).

Studies included in the meta-analyses

The mean responses measured as achieved 25(OH)D serum concentration in trial arms (both placebo/control and intervention groups) in a period of assumed minimal endogenous vitamin D synthesis were included in the preliminary analyses as long as the related individual trial arms met the following inclusion criteria:

Young and older adults as well as children – no pregnant women, no lactating women, no infants (following discussion with WG members, as these represent particular age/physiological conditions),
Vitamin D₃ only (as discussion with WG members suggested that intake of vitamin D₂ may have a different impact on 25(OH)D concentration),
Summary data available or possible to estimate/impute,
Dose of supplemented vitamin D ≤ 100 μg/day (Tolerable Upper Intake Level set by EFSA for adults (EFSA NDA Panel, 2012a)).

The inclusion criteria were applied at the arm level, as individual arms were considered the unit of analysis (except when mean differences were analysed).

After applying the inclusion criteria, 116 arms (49 trials) out of the 141 available in the contractor's data set (57 trials from 49 articles41) were left for the preliminary analyses (Appendix D.A, Table 8, third column showing the 25 excluded arms).

Upon evaluation of inconsistencies and outliers, a further 33 arms were excluded from the preliminary data set (Appendix D.A, Table 8 − fourth column). The final data set included 83 arms from 35 trials (Appendix D.B), of which four studies (nine arms) were carried out on children (overall age range: 2–17 years).

Absolute achieved mean values and mean differences were analysed to check for the inclusion of trials/arms in the dose–response analysis (preliminary meta-analyses) and to complement the results from the dose–response models (final meta-analyses; results reported below).

Achieved means from 83 arms (35 trials), also included in the final dose–response analysis, were displayed in forest plots with their 95% CIs and pooled weighted values estimated, both overall (pooled estimate: 57.9 nmol/L; 95% CI: 54.6–61.3) and by relevant subgroups (Appendix D.C, Figures 8–15).

Mean differences in achieved mean serum 25(OH)D concentration were calculated for 30 RCTs, out of the final 35 studies included in the dose–response analysis, where a control/placebo group and at least one intervention group were available (i.e. 5 trials out of 35 did not have a control group42). In case of multiple intervention groups, the achieved mean serum 25(OH)D of the first intervention arm (with the lowest dose) was selected to be compared to the achieved mean serum 25(OH)D of the control group. The pooled weighted mean difference across the 30 trials was 29.3 nmol/L (95% CI: 26.4–32.3) (Appendix D.D, Figure 16), with average achieved means of 41.3 nmol/L (SD = 10.3) and 70.8 nmol/L (SD = 14.1) in the control and intervention groups, respectively, and very close average baseline means (50.4 and 51.1 nmol/L, SD = 16). Analysis of weighted pooled estimate of mean differences in achieved mean serum 25(OH)D by 5 μg increase in total vitamin D intake (between 5 and 50 μg/day) is also reported in Appendix D.D (Figure 17).

Results from studies on specific populations (infants, lactating and pregnant women) were not included in separated meta-analyses (Table 8 of Appendix D.A) because their number (two arms on pregnant women, three arms on lactating women, three arms on infants) and characteristics were not deemed suitable (a minimum of three per subpopulation is requested); their results are addressed narratively in the contractor's report (Brouwer-Brolsma et al., 2016).

8 Meta-regression of the response of serum 25-hydroxyvitamin D to total vitamin D intake

Weighted linear meta-regression analyses of total vitamin D intake (i.e. habitual intake of the vitamin plus the supplemental dose) versus mean achieved serum or plasma 25(OH)D concentration measured at the end of the winter sampling points were performed.

The models were developed applying a random-effects approach ('random-effects meta-regression'), in which the extra variability due to heterogeneity is incorporated in the same way as in a random-effects meta-analysis, where the influence of more precise studies on the relationship is mitigated by the consideration of variability across studies. The approach allowed for extra residual heterogeneity among dose–response estimates not modelled by the explanatory variables identified and tested.

8.1 Studies included in the dose–response analysis

Meta-regression analyses were performed on the final data set (83 arms, 35 trials), as identified in Section 7 of this Appendix.

Most of the exclusions from the preliminary data set were based on inconsistencies in achieved means, mean differences (between intervention and control in the same trial) and net mean changes (between baseline and achieved mean in the same arm) of serum 25(OH)D (in the same trial across intervention groups and/or across trials in the same dose group). Careful reconsideration of study characteristics (e.g. design, type of participants, supplementation scheme, reporting issues, and summary data type) was the basis as to whether confirm exclusion of the identified arms (or entire related trial) (Appendix D.A, Table 8 – fourth column).

In addition, four arms were excluded based on model checking results (statistical outliers), after revision of all standardised residuals that were found to be either smaller than −2 or larger than +2. Two further exclusions were applied after reconsideration of the maximum supplemented vitamin D dose to be included, i.e. 50 μg/day, in order to model total vitamin D intakes that were not exceeding 100 μg/day (the UL set by EFSA) (Appendix D.A, Table 8 – fourth column).

8.2 Model construct

Two different model constructs of the dose–response relationship between plasma/serum 25(OH)D and total vitamin D intake were explored:

Log-linear: total vitamin D intake was transformed to the natural log (ln) before regression analysis; the regression intercept was set to 0 nmol of mean achieved 25(OH)D serum level to prevent negative values (which are biologically implausible). The intercept of the final adjusted model was not statistically significantly different from zero.

Linear: mean achieved serum 25(OH)D concentrations were regressed to total vitamin D intake on its original scale; the total vitamin D intake data points modelled were limited by a maximum intake dose of 35 μg/day, on the basis of evidence showing that the slope response of serum 25(OH)D to increasing dose becomes constant at such dose, as suggested by others (Aloia et al., 2008).

A non-linear response of serum 25(OH)D to vitamin D intake was expected due to metabolic kinetics (Heaney et al., 2008); in fact, the response of serum 25(OH)D is not best described by a linear fit model at doses above 35 μg/day.

The interest in exploring the linear model construct as an alternative to the curvilinear one was that the latter has a steep decline in achieved serum 25(OH)D concentrations particularly at the lower end of the range of total vitamin D intakes, and at zero intake the achieved serum 25(OH)D is forced to be 0 nmol/L to avoid a negative predicted value.

The WG decided to retain the log linear construct to better describe the dose–response shape and to be able to include results from higher dose trials (i.e. up to 50 μg/day).

8.3 Model fitting

For each random-effects meta-regression model, the statistics T ² (tau-squared, between-study variance) and Adjusted R ² were calculated. T ² was estimated using the restricted maximum likelihood method (Thompson and Sharp, 1999) with Knapp–Hartung modification of the estimates of the variance–covariance matrices of the regression coefficients (Knapp and Hartung, 2003) to reduce false-positive rates.

The change in T ² after inclusion of each covariate gives the amount of heterogeneity explained by the fitted model, and this value over the T ² from the null model gives the proportion of between-study variance explained (Adjusted R ² ).

T ² decreased from 312 to 46 in the final model, with included factors explaining up to 85% of heterogeneity (Appendix D.E, Table 13), i.e. ((312 − 46)/312) × 100 = 85% (Adjusted R ²) of between-study variance explained and 15% of unexplained heterogeneity.

The residual I ² statistics gives a measure of the percentage of the residual variation (the one not explained by the covariates) that is attributable to between-study heterogeneity.

Residual I² also decreased after inclusion of the final set of covariates, yet remaining quite high (87%) (Appendix D.E, Table 13).

In addition to the evaluation of the relative reduction of T ² and of the joint testing (using the F distribution) of covariates as introduced in the model, a backward elimination process was used to check the set of explanatory variables identified, by manual fitting in the final model, as significant predictors of the mean achieved serum concentrations.

8.4 Baseline measurements

The influence of the mean baseline 25(OH)D concentration on the dose–response relationship was described by plotting its values against the corresponding achieved mean values and explored in subgroup analyses (Appendix D.C, Figure 6; ≤ versus > 50 nmol/L) and meta-regression models (continuous covariate, Table 5 of this Appendix C). Bubble plots of net values (achieved 25(OH)D concentrations minus baseline values) were also considered to complement the dose–response analysis (not shown in this report).

After total vitamin D intake, the mean baseline 25(OH)D concentration was the factor explaining the highest proportion of between-study variability (17% in the simple meta-regression model – not shown in this report).

This is not surprising as it is likely that baseline values can serve as a surrogate for many influencing factors, potentially including some of those that could not be measured in the analysed trials. In fact, in the final adjusted model, the regression coefficient for the mean baseline was only marginally changed by the mutual adjustment for all the other included covariates (0.53 vs 0.48, (Appendix D.E, Table 13)).

8.5 Inter-individual variability on dietary intake

Previous analyses on vitamin D intake-status have encountered difficulties in taking into account the inter-individual variability on intake required to reach a chosen serum 25(OH)D cut-off.

The CI in meta-regression analyses provides an estimate of the uncertainty about the fitted response line due to sampling, but does not provide any estimate of the variability between individuals in terms of dietary intake of vitamin D needed to achieve a serum 25(OH)D concentration.

Attempts have been made to augment the meta-analytic approach by using individual data from vitamin D RCTs (Cashman et al., 2011a), which was not possible in the case of the current analysis as no individual data were available (Section 5.3.1 of the Scientific Opinion).

8.6 Model checking diagnostics

Outliers and influential studies were detected and tests for normality and homoscedasticity carried out to check for model assumptions (e.g. normality of the random effects).

The normal probability plot of the standardised predicted random effects did not show substantial departure from normality; outliers were identified by evaluation of standardised residual values smaller than −2 or larger than +2 (Appendix D.A, Table 8, fourth column) as estimated from the final models.

When several covariates are used in meta-regression, either in several separate simple meta-regressions or in one multiple meta-regression, there is an increased chance of at least one false-positive finding (type I error). The statistics obtained from the random permutations can be used to adjust for such multiple testing by comparing the observed t statistic for every covariate with the largest t statistic for any covariate in each random permutation (Higgins and Thompson, 2004).

Permutation-based p-values were calculated by running a Monte Carlo permutation test.

8.7 Dose–response influencing factors, investigation of heterogeneity between studies

A number of factors potentially influencing the dose–response relationship were identified a priori both from the relevant literature and upon feedback from the WG.

The following list was prioritised based on the outcome of WG's discussions; a selection of priority study-level characteristics was tested in independent subgroup analyses and incorporated in the meta-regression models one at a time and in the final multivariable model:

Total vitamin D intake: as continuous, as categorical (cut-offs determined by an increment of 5 μg/day; Appendix D.C, Figure 7),
Baseline serum concentration: as continuous, as dichotomous (cut-offs: 30 nmol/L (not shown in this report) and 50 nmol/L (Appendix D.C, Figure 6),
Study duration: ≤ 3 months vs > 3 months,
Latitude: as categorical, stratified by > 40°N to < 50°N and ≥ 50°N and 78°S,43
Assay method used: HPLC and LC–MS versus immunoassays (i.e. RIA, CBPA, ELISA),
Period of study publication: also related to trends in analytical methods (cut-off: year 2000) (not shown in this report),
Body Mass Index: a 'proxy' for body composition (which is not reported in the included trials); as continuous (study-level mean BMI), as per four categories: 'Normal weight', 'Overweight', 'Obese', 'Not reported' (Appendix D.C, Figure 14),
Ethnicity: a 'proxy' for skin pigmentation and some lifestyle habits that were usually not reported in the included trials; as per four categories: 'Caucasian', 'African', 'Mixed', 'Not Reported',
Co-supplemented calcium: as categorical (Yes, No/Unknown) (not shown in this report)
Funding source: as categorical ('Non-profit', 'Profit', 'Mixed', 'Not reported') (not shown in this report),
Age: as continuous (study-level mean age), as categorised according to three population groups (children, adults, older adults; the latter from trials where the reported or estimated mean age was ≥ 60 years) (Appendix D.C, Figure 14)
Sex: as categorical based on percentage of males ('Both' for studies on mixed populations, 'Women' for studies on women only, 'Men' for studies on men only) (Appendix D.C, Figure 15)
Risk of bias (RoB) dimensions: all individually categorised as 'Yes', 'No/Unknown' (adequate randomisation, adequate allocation concealment, adequate blinding description, compliance assessed, drop-outs addressed, dose check reported); as combined by the contractor (Brouwer-Brolsma et al., 2016) in an overall RoB assessment ('High', 'Moderate', 'Low' RoB) (Appendix D.B, Table 12).

The following further categorisations were also applied and tested a posteriori:

Duration: ≤ 3 months vs > 3 months & < 6 months vs 1–2 years (Appendix D.C, Figure 8),
Latitude: < 50°N, 50–55°N, > 55°N (Appendix D.C, Figure 9). For 76% of arms, latitude was > 50°N (Table 9 in Appendix D.B),
Assay method used: RIA versus HPLC versus LC–MS versus CPBA versus ELISA & Not Reported versus Other30 (Appendix D.C, Figure 11). In the final model (Section 8.8), each analytical method was retained as an individual category to be able to estimate the specific effects,
Ethnicity: 'Caucasian', 'Mixed', 'Not Reported'. 'African' was grouped to the 'Mixed' category, as it included three arms only (Appendix D.C, Figure 12).

Study start period was subsequently considered instead of publication year as a better proxy to the temporal trends in assay method use (as continuous − since year of first study in analysis, i.e. 1985; as dichotomous -before or after 2000) (Appendix D.C, Figure 10).

Pooled estimates in the placebo/control arms and intervention arms were also reported for descriptive purposes (Appendix D.C, Figure 5).

All results (Appendix D.C, Figures 4–15) were interpreted only qualitatively and group summary estimates compared by visual inspection; subgroup comparisons are observational in nature and results from statistical testing should not be used to infer that estimates differ from one stratum to another.

8.8 Derivation of DRVs

The meta-regression analysis carried out on the selected arms resulted in two predictive equations of achieved serum 25(OH)D:

y = 23.2 ln (total vitamin D intake) (unadjusted model) (Appendix D.F, Figure 18) and

y = 16.3 ln (total vitamin D intake) adjusted for baseline concentration (continuous; μg/day), latitude (continuous; °N), study start year (continuous; years since first study in analysis – 1985), type of analytical method applied (RIA, HPLC, LC-MS, CPBA, ELISA/not reported, Other), assessment of compliance (yes, no/unknown) (Table 5 of this Appendix C, and Figure 19 of Appendix D.F).

Age and sex were not included in the final model as they did not explained further neither the within- nor the between- study variability. The role of BMI was also tested in the subset of arms for which such information was available (83%); overweight and obese subgroups from the study populations showed on average higher achieved means when compared to the normal weight group (Appendix D.C, Figure 13), but lower values once adjusted for all other covariates. BMI was not included in the final model as it did not reach statistical significance in the preliminary analyses from the preliminary data set (116 arms) and in consideration of potential ecological fallacy (i.e. associations with mean BMI values when available or calculated from mean height and mean weight at study-level are not necessarily consistent with associations with individual-level BMI values).

Table 5. Adjusted meta-regression model (outcome variable: mean achieved 25(OH)D in nmol/L; n = 83 arms)

Covariate	β coefficient	SE	P >t	95% CI
Ln of total vitamin D intake – μg/day	16.33	0.94	< 0.001	14.45–18.21
Mean Baseline 25(OH)D – nmol/L	0.50	0.05	< 0.001	0.39–0.61
Latitude – °N	− 0.46	0.09	< 0.001	−0.63 to −0.29
Study start year (years since 1985)	0.93	0.21	< 0.001	0.51–1.35
Assay
RIAa ^a Reference category.	0.00
HPLC	−1.93	3.29	0.56	−8.49 to 4.62
LC-MS	−4.72	3.00	0.12	−10.69 to 1.26
CPBA	0.63	3.86	0.87	−7.07 to 8.33
ELISA/nr	−6.40	2.68	0.02	−11.73 to −1.06
Other	1.30	3.61	0.72	−5.89 to 8.49
Compliance assessed
Yesa ^a Reference category.	0.00
No/unknown	7.79	2.97	0.01	1.86–13.71

CI: confidence interval; nr: not reported; SE: standard error.
P >t: indicates the probability of the hypothesis that the beta-coefficient = 0 (since p = 0.05 is conventionally assumed as the cut-off for statistical significance in the analysis, a p value lower than 0.05 provides good evidence that the beta-coefficient is significantly different from 0).
^a Reference category.

The same equations were used both to predict the achieved mean serum 25(OH)D levels conditional to total vitamin D intakes of 5, 10, 15, 20, 50, 100 μg/day (Table 6 of this Appendix C) and to estimate the total vitamin D intakes that would achieve serum 25(OH)D concentrations of 50, 40, 30, 25 nmol/L (Table 7 of this Appendix C).

All values were calculated by using the regression equations of the predicted mean, of the lower and upper limits of the 95% CI of the predicted mean and of the lower and upper limits of the 95% prediction interval (PI) of the predicted mean. In the adjusted multivariable models, all covariates were set to their mean values (mean baseline 25(OH)D: 50.7 nmol/L; latitude: 53°N; study start year: 2005; assay – HPLC: 10%; LC-MS: 18%; CPBA: 13%; ELISA: 20%; other: 8%; compliance not assessed/unknown: 27%).

A stratified analysis was carried out to quantify the impact of the exclusions of the four trials on children (nine arms) on the predicted achieved mean serum 25(OH)D levels (Appendix D.G, Table 14, ADULTS estimates) and estimated total vitamin D intakes (Appendix D.G, Table 15, ADULTS estimates). In the restricted data set (74 arms), there was an overall small decrease in all serum estimates (and consequently a small increase in total intakes that would achieve target values); this is possibly due both to the fact that 'children' arms were just 9 and that children tend to achieve the same concentrations as the adults at a lower total intake (Appendix D.G, Table 14, CHILDREN estimates). Overall estimates did not substantially change as compared to the full data set including children.

Values based only on the four children trials were not calculated in the fully adjusted meta-regressions, as they would have required a much higher minimum number of 'points' per covariate (at least 10 arms for each included factor); instead, values from a model adjusted by mean baseline 25(OH)D were provided. As such, these estimates are not directly comparable to the adults' ones, as they are not adjusted for the same set of covariates. The unadjusted model showed lower average intakes, but estimates were much less precise (with 95% CI overlapping with those from the adults data), and could only be evaluated qualitatively (Appendix D.G, Table 15, CHILDREN estimates).

In the meta-analytic context, when a random-effects approach is applied, the CI reflects the precision with which we estimate the pooled (across studies) mean effect size (via the available sample of studies), while the PI reflects the actual dispersion of the true effects around the mean effect size.

If, for instance, there is an estimated mean response of 50 with a CI of 40 to 60, the range of 40 to 60 includes with a certain frequency (conventionally 95% of the times) the true mean response in the population of studies from which the sample was drawn.

From a related PI of 30 to 70, it can be considered that probably (conventionally 95% of the times) such range will include the true effect in a new study from the same population of studies. If the number of studies were infinite, then the CI width would approach zero but the PI would show little change.

When interpreting the intervals drawn around the meta-regression lines, the CI illustrates the uncertainty about the position of the line (i.e. across-study conditional means), while the PI illustrates the uncertainty about the true mean effect that would be predicted in a future study (i.e. the dispersion of the true effects around their mean).

As such, it is possible to think of the latter only as an approximation of the interval that would allow for estimation of the requirements for 95% of the population, as it refers to the population of mean responses (not individual responses) as analysed in the random-effects model.

Table 6. Predicted achieved serum 25(OH)D at selected values of total vitamin D intake

Table 7. Estimated vitamin D intakes at selected serum 25(OH)D cut-off values

9 Quality of the body of evidence: addressing risk of bias

The rating by the contractor of individual trials in terms of RoB (individual dimensions and overall assessment) (Brouwer-Brolsma et al., 2016) was used to evaluate whether heterogeneity of results could be attributed to differences in internal validity, both in the meta-analyses and meta-regression models (Appendix D.B, Table 12). The following approaches were discussed and applied accordingly:

To run the analysis on low-moderate-risk trials only (restriction): this option could not be applied as the proportion of low-risk arms was only 16% (plus moderate-risk ones accounting for an additional 18%). The trade-off between bias and precision would have been too much towards (possibly) more valid but less precise estimates;
To run a sensitivity analysis and see how the response changes if high-risk studies are excluded: this was not carried out considering that the majority of trials were rated high-RoB;
To run a subgroup analysis (or meta-regression) re-grouping the RoB variable into a dichotomous one: this was considered but the covariate was tested as originally coded (low, moderate, high risk). The lack of a statistically significant difference between studies at high and low RoB (data not shown in this report) should be interpreted cautiously as meta-regression analyses are observational in nature;
To use individual dimensions as recorded by the contractor (Brouwer-Brolsma et al., 2016): each RoB dimension was evaluated in univariate and multivariable analyses. Assessed compliance (categorised as yes versus no/unknown and independently of its definition across trials) was found to play a role in further explaining the variability between studies (Appendix D.E, Table 13); all other dimensions (randomisation appropriate, allocation concealment, etc.) were not statistically significantly impacting on the estimates (not shown in this report);
To integrate a qualitative (narrative) evaluation of RoB in the discussion of the analysis results.

10 Sensitivity analyses

A number of sensitivity analyses were carried out to evaluate whether the findings were robust to the assumptions made in the systematic review protocol and the analyses (e.g. meta-regression models).

When sensitivity analyses show that the overall result and conclusions are not substantially affected by the different decisions that could be made during the review process, the results of the review can be regarded with a higher degree of certainty.

There were a number of assumptions/decisions/issues provisionally identified that could potentially be tested in sensitivity analyses by comparing the results obtained with alternative input parameters to those from the default model or by restricting to specific subsets; none of them raised serious concerns about the robustness of the overall analysis (the most substantial departures were detected in the smallest, then less representative, subsets of the final data set).

The following analysis were considered:

On data cleaning issues: implausible values, missing data,
On quality dimensions: compliance assessment,
On analytical approaches: data imputation; cut-off points, choice of categories,
On eligibility criteria: fortified food trials; range of doses (exclusion of doses higher than 100 μg/day); characteristics of participants (exclusion of non-healthy volunteers, of supplement users, etc.; Appendix D.H, Table 16).

11 Observational studies: contribution of their results to the analysis

Meta-analyses were performed separately for RCTs and observational studies (prospective cohort studies) on the basis that, in principle, evidence from randomised and non-randomised studies is not considered comparable. Eight prospective observational studies from seven articles were included. (Appendix D.I, Table 17). They represented 11 study groups (e.g. children versus adults in Andersen 2013, Caucasian group versus Asian group in Darling et al. (2013), Caucasian from one study centre versus a group of Caucasian and a group of Asian people in another study centre in MacDonald et al. (2011)), three of which were on children (mean age between 11 and 16 years).

Achieved mean serum 25(OH)D concentration (and 95% CI) was investigated by study group (Appendix D.I, Figure 20), as well as by relevant subgroups: age (children versus adults; Appendix D.I, Figure 21), baseline mean serum 25(OH)D concentrations (≤ versus > 50 nmol/L; Appendix D.I, Figure 22) and latitude (< 50°N versus ≥ 50°N; Appendix D.I, Figure 23).

12 Publication bias

Several systematic reviews of empirical studies have found that studies with statistically significant or positive results are more likely to be published than those with non-significant or negative results. Investigators' decisions not to submit papers with negative results for publication, rather than editors' rejection of such papers, tend to be the main source of publication bias. Studies with statistically significant results also tend to be published earlier than studies with non-significant results. If studies are missing from a systematic review for these reasons, effects may be over-estimated (Higgins and Green, 2011).

Publication bias was examined by inspecting funnel plots (Sterne and Egger, 2001) and by performing the Egger's test for funnel plot asymmetry (Egger et al., 1997) on mean differences in achieved mean serum 25(OH)D from the 30 RCTs included in the meta-analyses (see Section 8).

Egger's test performs a linear regression of the intervention effect estimates on their standard errors, weighting by 1/(variance of the intervention effect estimate) (Appendix D.J, Figure 24); the test was not statistically significant (p = 0.149).

Funnel plots investigate the association between study size and effect size; there was no particular indication of funnel plot asymmetry, as trials testing a dose of 5 to < 10 μg/day were missing in the right-hand side of the funnel while trials testing 45 μg/day and more were missing in the left-hand side (Appendix D.J, Figure 25).

13 Uncertainty analysis

Sources of uncertainty and their potential impact on the final estimates, where possible, were identified and discussed:

General interpretation of meta-regression results – the associations derived from meta-regressions are observational and have a weaker interpretation than those derived from randomised comparisons; this applies especially when population characteristics are included as means at study level,
Inter-individual variability on intake – failure to account for it may lead to underestimation of the predicted intake of vitamin D needed to maintain a specified serum 25(OH)D concentration (Cashman et al., 2011a),
Predicted achieved mean serum 25(OH)D concentrations and estimated total vitamin D intakes calculated based on the 95% CI of the predicted mean from the adjusted models were less accurate than those from the unadjusted ones, due to the approximation of the fitting on the pair wise limits,
Predictions from the lower range of the total vitamin D intakes are less accurate than those for higher values because of the log-linear construct (not optimal fitting in that intake range),
Ecological fallacy – key risk factors that vary across populations and that can be measured only as aggregate values, such as age, gender and BMI, are difficult to address adequately by meta-regression. One reason for this is that aggregated values tend to exhibit little between-study variation, thus providing minimal information across the potential range of the factor. Use of aggregated values may also introduce bias because of the failure to account for the within-study variation (Thompson and Higgins, 2002),
Selection of RCTs/arms – the main objective of the additional exclusion of arms from the final data set (Appendix D.A, Table 8) was to try to 'remove' as much heterogeneity as possible that could be attributable to differences in design, bias, and/or methods, so that only 'clinical' heterogeneity (i.e. between-study variability due to population's features) would be left to be modelled and characterised. It is difficult to quantify the potential relative misclassification due to such a selection; the proportion of heterogeneity explained by the influencing factors in the final subset was higher than that in the preliminary data set (85% vs 56%), but the regression coefficients of all covariates were almost unchanged. This could be interpreted as a relative reduction of heterogeneity more in its methodological component across included studies, due to the nature of the criteria applied for the additional exclusions.

Appendix D – Dose–response analysis undertaken by EFSA of serum 25(OH)D to total vitamin D intake: appendices

A. List of trials arms not included in the meta-analyses and dose–response analysis

Table 8. Reasons for exclusions from preliminary data set and final data set (total: 58 arms excluded out of 141)

RCT arms	Suppl. vitamin D dose (μg/day)	Reasons for exclusion from preliminary data set (25 arms excluded out of 141)	Reasons for exclusion from final data set (33 additional arms excluded)
Ala-Houhala et al. (1986)aa ^a e.g. (Ala-Houhala et al., 1986)a, (Ala-Houhala et al., 1986)b and (Ala-Houhala et al., 1986)c (as cited in Brouwer-Brolsma et al. (2016)) refer to the same study, but different population groups (e.g. in this case: pregnant women, lactating women and infants).	12.5	Study on pregnant women	–
Ala-Houhala et al. (1986)a	0	Study on pregnant women	–
Ala-Houhala et al. (1986)ba ^a e.g. (Ala-Houhala et al., 1986)a, (Ala-Houhala et al., 1986)b and (Ala-Houhala et al., 1986)c (as cited in Brouwer-Brolsma et al. (2016)) refer to the same study, but different population groups (e.g. in this case: pregnant women, lactating women and infants).	50	Study on lactating women	–
Ala-Houhala et al. (1986)b	25	Study on lactating women	–
Ala-Houhala et al. (1986)b	0	Study on lactating women	–
Ala-Houhala et al. (1986)ca ^a e.g. (Ala-Houhala et al., 1986)a, (Ala-Houhala et al., 1986)b and (Ala-Houhala et al., 1986)c (as cited in Brouwer-Brolsma et al. (2016)) refer to the same study, but different population groups (e.g. in this case: pregnant women, lactating women and infants).	10	Study on infants	–
Ala-Houhala et al. (1988b)	10	Study with supplemented vitamin D₂	–
Ala-Houhala et al. (1988b)	0	Study with supplemented vitamin D₂	–
Atas et al. (2013)	10	Study on infants	–
Atas et al. (2013)	5	Study on infants	–
Barger-Lux et al. (1998)	1,250	Arm with supplemented dose > 100 μg/day	–
Barger-Lux et al. (1998)	250	Arm with supplemented dose > 100 μg/day	–
Brazier et al. (2002)	20	–	Methodological considerationsc ^c 'Methodological considerations' in this table: design, type of participants, supplementation scheme, reporting issues, and summary data type (see Appendix C). applicable to whole study
Brazier et al. (2002)	0	–	Inconsistent net mean change + methodological considerations
Close et al. (2013a)	125	Arm with supplemented dose > 100 μg/day	–
Close et al. (2013a)	0	–	Inconsistent net mean change and achieved mean + methodological considerations
Forman et al. (2013)	100	–	Arm with supplemented dose ≥ 100 μg/day
Heaney et al. (2003b)	250	Arm with supplemented dose > 100 μg/day	–
Heaney et al. (2003b)	125	Arm with supplemented dose > 100 μg/day	–
Holick et al. (2008)	25	Arm with supplemented vitamin D₂	–
Holick et al. (2008)	25	Arm with supplemented vitamin D₂	–
Holm et al. (2008)	5	–	Supplementation scheme was 5 μg/3 days + inconsistent mean difference
Holm et al. (2008)	0	–	Control group only left from study
Honkanen et al. (1990)b	45	–	Methodological considerations applicable to whole study
Honkanen et al. (1990)b	0	–	Statistical outlier
Johnson et al. (2005)	15	–	Inconsistent achieved mean + methodological considerations
Johnson et al. (2005)	0	–	Methodological considerations applicable to whole study
Johnson et al. (2005)	0	–	Methodological considerations applicable to whole study
Larsen et al. (2012)	25	–	Statistical outlier
Larsen et al. (2012)	0	–	Control group only left from study
Lehmann et al. (2013)	50	Arm with supplemented vitamin D₂	–
Mocanu et al. (2009)	125	Study with supplemented dose > 100 μg/day	–
Nelson et al. (2009)	20	–	Methodological considerations applicable to whole study
Nelson et al. (2009)	0	–	Inconsistent net mean change + methodological considerations
Patel et al. (2001)a	20	–	Inconsistent achieved mean + methodological considerations
Patel et al. (2001)a	0	–	Methodological considerations applicable to whole study
Patel et al. (2001)b	20	–	Inconsistent achieved mean + methodological considerations
Porojnicu et al. (2008)	5	Quantitative data on response not available	–
Porojnicu et al. (2008)	0	Quantitative data on response not available	–
Rich-Edwards et al. (2011)	7.5	–	Statistical outlier (fortified UHT milk arm)
Schmidt and Zirkler (2011)	5	–	Inconsistent mean difference + methodological considerations
Schmidt and Zirkler (2011)	0	–	Control group only left from study
Sorva et al. (1994)	25	Arm with supplemented vitamin D₂	–
Sorva et al. (1994)	25	–	Statistical outlier
Sorva et al. (1994)	0	–	Control group only left from study
Vieth et al. (2001)	100	–	Arm with supplemented dose ≥ 100 μg/day
White et al. (2009)	3	Mixed interventionb ^b Food fortified with vitamin D + training exercise, compared to supplements without vitamin D +training exercise. , very high baseline values	–
White et al. (2009)	0	Mixed interventionb ^b Food fortified with vitamin D + training exercise, compared to supplements without vitamin D +training exercise. , very high baseline values	–
White et al. (2009)	0	Mixed interventionb ^b Food fortified with vitamin D + training exercise, compared to supplements without vitamin D +training exercise. , very high baseline values	–
Wood et al. (2014)_nw	25	–	Methodological considerations applicable to whole study
Wood et al. (2014)_nw	10	–	Methodological considerations applicable to whole study
Wood et al. (2014)_nw	0	–	Inconsistent baseline mean value + methodological considerations
Wood et al. (2014)_ow	25	–	Methodological considerations applicable to whole study
Wood et al. (2014)_ow	10	–	Methodological considerations applicable to whole study
Wood et al. (2014)_ow	0	–	Inconsistent baseline mean value + methodological considerations
Wood et al. (2014)_ob	25	–	Methodological considerations applicable to whole study
Wood et al. (2014)_ob	10	–	Methodological considerations applicable to whole study
Wood et al. (2014)_ob	0	–	Inconsistent baseline mean value + methodological considerations

nw: normal weight; ob: obese; ow: overweight; UHT: Ultra-high temperature.
Initial data set: 141 arms, preliminary data set: 116 arms (after exclusion if 25 arms), final data set: 83 arms (after additional exclusion of 33 arms), total of excluded arms: 58 (i.e. 25 + 33).
^a e.g. (Ala-Houhala et al., 1986)a, (Ala-Houhala et al., 1986)b and (Ala-Houhala et al., 1986)c (as cited in Brouwer-Brolsma et al. (2016)) refer to the same study, but different population groups (e.g. in this case: pregnant women, lactating women and infants).
^b Food fortified with vitamin D + training exercise, compared to supplements without vitamin D +training exercise.
^c 'Methodological considerations' in this table: design, type of participants, supplementation scheme, reporting issues, and summary data type (see Appendix C).

B. Trials included in the dose–response analysis (35 trials) – main study characteristics

Table 9. Country, latitude, age, sex, duration (35 trials)

Source	Country	Latitude (°N)	Mean age (years)	Age range (years)	Males (%)	Duration (weeks)
Barger-Lux et al. (1998)	USA	41.2	28	20–37	100	8
Barnes et al. (2006)	IE	54.8	22	18–27	50	8
Bischoff-Ferrari et al. (2003)	CH	47.3	85	–	0	12
Bolton-Smith et al. (2007)	UK	56.3	70	60+	0	104
Bonjour et al. (2013)	FR	50.7	86	60+	0	8
Braam et al. (2003)	NL	50.9	55	50–60	0	156
Cashman et al. (2008)	IE	51	30	20–40	50	22
Cashman et al. (2009)	IE	51	71	64+	40	22
Cashman et al. (2012)	IE	51	57	50+	38	10
Cashman et al. (2014)	IE	51	60	50+	28	15
de Gruijl and Pavel (2012)	NL	52.2	24	18–30	9	8
DeLappe et al. (2006)	IE	53.2	80	–	0	13
Forman et al. (2013)	USA	42.2	51	30–79	35	13
Goussous et al. (2005)	USA	42.2	65	50+	27	13
Hansen et al. (2010)	NO	60.4	35	20–60	100	23
Harris and Dawson-Hughes (2002)a	USA	42	26	18–35	100	8
Harris and Dawson-Hughes (2002)b	USA	42	70	62–79	100	8
Heaney et al. (2003b)	USA	41.2	39	–	100	20
Heikkinen et al. (1998)	FI	62.9	51	47–56	0	52
Holick et al. (2008)	USA	42.3	60	18–84	31	6
Honkanen et al. (1990)a	FI	63	70	67–72	0	11
Hower et al. (2013)	DE	51.2	4	2–6	56	20
Keane et al. (1998)	IE	53.2	78	65–92	24	47
Lehmann et al. (2013)	DE	51.47	43	19–67	33	8
Madsen et al. (2013)a	DK	55.7	10	4–17	48	26
Madsen et al. (2013)b	DK	55.7	36	18–60	50	26
Meier et al. (2004)	DE	50	54	33–78	33	25
O'Connor et al. (2010)	DK	55.4	11	11–12	0	52
Pekkarinen et al. (2010)	FI	61	74	69–79	0	52
Rich-Edwards et al. (2011)	MN	48	10	9–11	53	7
Smith et al. (2009)	AQ	78a ^a Latitude of 78°S.	43	–	75	22
Trautvetter et al. (2014)	DE	50.6	42	–	40	8
Vieth et al. (2001)	CA	43	41	–	33	8
Viljakainen et al. (2006b)	FI	61	71	65–85	0	12
Viljakainen et al. (2009)	FI	61	29	21–49	100	26

AQ: Antarctica; CA: Canada; CH: Switzerland; DE: Germany; DK, Denmark; FI: Finland; FR: France; IE: Ireland; MN: Mongolia; NL: the Netherlands; NO: Norway; UK: United Kingdom; USA: United States of America.
e.g. (Madsen et al., 2013)a and (Madsen et al., 2013)b (as cited in Brouwer-Brolsma et al. (2016)) refer to the same study, but different population groups (e.g. in this case: children and adults).
^a Latitude of 78°S.

Table 10. Start year, funding, ethnicity, analytical method, Ca co-supplementation (35 trials)

Source	Start year	Funding	Ethnicity	Analytical method	Ca co-suppl.
Barger-Lux et al. (1998)	1997	Mixed	Mixed	HPLC	No/unknown
Barnes et al. (2006)	2005	–	–	ELISA	Yes
Bischoff-Ferrari et al. (2003)	1999	Mixed	–	RIA	Yes
Bolton-Smith et al. (2007)	2003	Mixed	–	RIA	Yes
Bonjour et al. (2013)	2010	Profit	–	ELISA	Yes
Braam et al. (2003)	1997	Mixed	Caucasian	RIA	Yes
Cashman et al. (2008)	2006	Non-profit	Caucasian	ELISA	No/unknown
Cashman et al. (2009)	2007	Non-profit	Caucasian	ELISA	No/unknown
Cashman et al. (2012)	2011	Mixed	Caucasian	ELISA	No/unknown
Cashman et al. (2014)	2012	Non-profit	Caucasian	LC-MS	No/unknown
de Gruijl and Pavel (2012)	2010	Mixed	Mixed	RIA	No/unknown
DeLappe et al. (2006)	2003	–	–	RIA	Yes
Forman et al. (2013)	2007	Mixed	African	RIA	Yes
Goussous et al. (2005)	2003	Mixed	Mixed	RIA	Yes
Hansen et al. (2010)	2008	Non-profit	Mixed	RIA	No/unknown
Harris and Dawson-Hughes (2002)a	2000	Mixed	–	CPBA	No/unknown
Harris and Dawson-Hughes (2002)b	2000	Mixed	–	CPBA	No/unknown
Heaney et al. (2003b)	2001	Non-profit	–	Other	No/unknown
Heikkinen et al. (1998)	1990	Mixed	–	CPBA	Yes
Holick et al. (2008)	2007	Mixed	Mixed	LC-MS	No/unknown
Honkanen et al. (1990)a	1985	Mixed	–	CPBA	Yes
Hower et al. (2013)	2010	Profit	Caucasian	Other	No/unknown
Keane et al. (1998)	1993	Profit	–	CPBA	No/unknown
Lehmann et al. (2013)	2012	Non-profit	–	LC-MS	No/unknown
Madsen et al. (2013)a	2010	Mixed	–	LC-MS	No/unknown
Madsen et al. (2013)b	2010	Mixed	–	LC-MS	No/unknown
Meier et al. (2004)	2002	–	–	RIA	Yes
O'Connor et al. (2010)	2008	Non-profit	Mixed	HPLC	No/unknown
Pekkarinen et al. (2010)	2006	Non-profit	Caucasian	HPLC	Yes
Rich-Edwards et al. (2011)	2009	Mixed	Mixed	LC-MS	No/unknown
Smith et al. (2009)	2007	Non-profit	Caucasian	RIA	No/unknown
Trautvetter et al. (2014)	2011	Profit	–	ELISA	Yes
Vieth et al. (2001)	2000	Profit	Mixed	RIA	No/unknown
Viljakainen et al. (2006b)	2002	Non-profit	–	HPLC	No/unknown
Viljakainen et al. (2009)	2007	Non-profit	Caucasian	Other	No/unknown

Ca co-suppl: calcium co-supplementation; CPBA: competitive protein binding assay; ELISA: enzyme-linked immunosorbent assay; HPLC: high-performance liquid chromatography; LC-MS: liquid chromatography-mass spectroscopy; RIA: radioimmunoassay.
e.g. (Madsen et al., 2013)a and (Madsen et al., 2013)b (as cited in Brouwer-Brolsma et al. (2016)) refer to the same study, but different population groups (e.g. in this case: children and adults).

Table 11. Vitamin D intakes, summary data (mean response with standard deviation) and body mass index (BMI) (35 trials, 83 arms)

Source	Habitual vitamin D intake (μg/day)	Supplemental Vitamin D dose (μg/day)	Total vitamin D intake (μg/day)	Participants per arm (n)	Baseline Mean 25(OH)D (nmol/L)	Baseline 25(OH)D SD (nmol/L)	Achieved Mean 25(OH)D (nmol/L)	Achieved 25(OH)D SD (nmol/L)	Mean BMI (kg/m²)
Barger-Lux et al. (1998)a ^a Trials for which habitual dietary intake was imputed from national survey data (age-, sex-specific).	5	25	30.0	13	67	25	96	18	25.7
Barnes et al. (2006)	1.6	15	16.6	12	48	16	87	25	24.8
Barnes et al. (2006)	2.4	0	2.4	15	56	19	48	17	22.9
Bischoff-Ferrari et al. (2003)a ^a Trials for which habitual dietary intake was imputed from national survey data (age-, sex-specific).	3.3	20	23.3	62	36	24	66	25	24.7
Bischoff-Ferrari et al. (2003)	3.3	0	3.3	60	35	24	32	12	24.7
Bolton-Smith et al. (2007)	5.9	10	15.9	49	62	17	71	16	26.1
Bolton-Smith et al. (2007)	5.6	10	15.6	50	62	15	74	15	25.8
Bolton-Smith et al. (2007)	5	0	5.0	56	57	15	49	13	26.2
Bonjour et al. (2013)a ^a Trials for which habitual dietary intake was imputed from national survey data (age-, sex-specific).	2.8	10	12.8	29	19	5	45	16	26.2
Bonjour et al. (2013)	2.8	0	2.8	27	16	5	21	16	26.6
Braam et al. (2003)a ^a Trials for which habitual dietary intake was imputed from national survey data (age-, sex-specific).	3.2	8	11.2	56	57	18	62	15	25.1
Braam et al. (2003)	3.2	8	11.2	46	56	14	62	11	25.5
Braam et al. (2003)	3.2	0	3.2	60	51	14	56	13	26.1
Cashman et al. (2008)	3.6	15	18.6	53	74	25	71	19	26.1
Cashman et al. (2008)	3.5	10	13.5	57	73	27	60	14	26.1
Cashman et al. (2008)	4.3	5	9.3	48	67	31	52	11	26.1
Cashman et al. (2008)	3.4	0	3.4	57	73	27	39	13	26.1
Cashman et al. (2009)	4.8	15	19.8	48	55	23	75	21	28.9
Cashman et al. (2009)	4.2	10	14.2	53	56	22	70	18	28.9
Cashman et al. (2009)	4.1	5	9.1	48	55	23	56	18	28.9
Cashman et al. (2009)	4.7	0	4.7	55	61	27	42	21	28.9
Cashman et al. (2012)	7.6	20	27.6	13	50	16	69	9	28.3
Cashman et al. (2012)	6.5	0	6.5	16	43	13	41	11	28.3
Cashman et al. (2014)	4.4	20	24.4	27	54	25	80	19	26.7
Cashman et al. (2014)	4.4	0	4.4	28	58	17	42	15	26.7
Cashman et al. (2014)	4.4	20	24.4	34	54	22	74	15	26.7
Cashman et al. (2014)	4.4	0	4.4	32	54	17	41	16	26.7
de Gruijl and Pavel (2012)a ^a Trials for which habitual dietary intake was imputed from national survey data (age-, sex-specific).	2.7	25	27.7	37	58	18	93	20	22.4
de Gruijl and Pavel (2012)	2.7	0	2.7	33	62	24	55	21	22.3
DeLappe et al. (2006)a ^a Trials for which habitual dietary intake was imputed from national survey data (age-, sex-specific).	3.4	20	23.4	51	42	27	60	27	–
Forman et al. (2013)a ^a Trials for which habitual dietary intake was imputed from national survey data (age-, sex-specific).	4.5	50	54.5	65	36	24	87	24	31
Forman et al. (2013)	4.5	25	29.5	56	41	22	74	22	31
Forman et al. (2013)	4.5	0	4.5	64	41	24	38	24	31
Goussous et al. (2005)	3.8	20	23.8	23	49	17	66	15	26.7
Goussous et al. (2005)	4.6	20	24.6	29	48	16	64	16	30.9
Hansen et al. (2010)a ^a Trials for which habitual dietary intake was imputed from national survey data (age-, sex-specific).	6.7	7	13.7	15	48	15	60	16	–
Hansen et al. (2010)	6.7	1	7.7	14	48	25	49	20	–
Harris and Dawson-Hughes (2002)a	1.8	20	21.8	13	60	16	82	12	25
Harris and Dawson-Hughes (2002)a	3.3	0	3.3	12	49	17	44	17	25.1
Harris and Dawson-Hughes (2002)b	3.5	20	23.5	14	62	16	84	19	29
Harris and Dawson-Hughes (2002)b	1.5	0	1.5	11	54	18	49	18	30
Heaney et al. (2003b)a ^a Trials for which habitual dietary intake was imputed from national survey data (age-, sex-specific).	5.4	25	30.4	17	72	16	80	16	26.2
Heaney et al. (2003b)	5.4	0	5.4	16	70	24	60	24	26.2
Heikkinen et al. (1998)a ^a Trials for which habitual dietary intake was imputed from national survey data (age-, sex-specific).	8.2	7.5	15.7	17	28	12	38	8	24.8
Heikkinen et al. (1998)	8.2	7.5	15.7	18	24	8	33	8	25.7
Heikkinen et al. (1998)	8.2	0	8.2	18	28	13	25	8	24.7
Holick et al. (2008)a ^a Trials for which habitual dietary intake was imputed from national survey data (age-, sex-specific).	4.4	25	29.4	20	49	28	65	28	30
Holick et al. (2008)	4.4	0	4.4	10	47	22	45	22	29.3
Honkanen et al. (1990)aa ^a Trials for which habitual dietary intake was imputed from national survey data (age-, sex-specific).	8.7	45	53.7	25	43	17	81	13	–
Honkanen et al. (1990)a	8.7	0	8.7	26	36	12	23	12	–
Hower et al. (2013)	1.9	7.1	9.0	39	67	25	65	24	–
Hower et al. (2013)	1.9	0.1	2.0	24	58	22	44	19	–
Keane et al. (1998)a ^a Trials for which habitual dietary intake was imputed from national survey data (age-, sex-specific).	3.6	5	8.6	24	24	5	46	11	–
Keane et al. (1998)	3.6	0.1	3.7	18	25	5	32	14	–
Lehmann et al. (2013)a ^a Trials for which habitual dietary intake was imputed from national survey data (age-, sex-specific).	3.2	50	53.2	42	44	23	89	22	23.7
Lehmann et al. (2013)	3.2	0	3.2	19	41	15	32	13	23.7
Madsen et al. (2013)a	2.3	7.9	10.2	154	75	17	68	4	–
Madsen et al. (2013)a	2.2	0	2.2	167	76	20	43	5	–
Madsen et al. (2013)b	2.4	5.4	7.8	201	76	20	66	4	–
Madsen et al. (2013)b	2.2	0	2.2	204	73	22	41	6	–
Meier et al. (2004)	3.2	12.5	15.7	27	75	29	88	20	26.1
Meier et al. (2004)	3.2	0	3.2	16	77	23	51	21	26.2
O'Connor et al. (2010)a ^a Trials for which habitual dietary intake was imputed from national survey data (age-, sex-specific).	2.3	10	12.3	33	48	16	58	14	18.1
O'Connor et al. (2010)	2.3	0	2.3	34	48	18	40	18	18.1
Pekkarinen et al. (2010)	6.4	20	26.4	20	58	10	74	10	26.9
Rich-Edwards et al. (2011)b ^b Values for Rich-Edwards et al. (2011) were imputed from Madsen et al. (2013) (children with same mean age).	2.2	7.5	9.7	140	20	10	50	15	16.4
Rich-Edwards et al. (2011)	2.2	7.5	9.7	109	17	7	52	15	16.5
Rich-Edwards et al. (2011)	2.2	0	2.2	101	20	10	20	10	17
Smith et al. (2009)	8.9	50	58.9	18	45	14	71	23	28
Smith et al. (2009)	8.2	25	33.2	19	44	19	63	25	31
Smith et al. (2009)	7.6	10	17.6	18	44	18	57	15	29
Smith et al. (2009)	15.7	0	15.7	7	36	17	34	12	28
Trautvetter et al. (2014)	6.2	10	16.2	20	46	20	70	20	25
Trautvetter et al. (2014)	6.5	10	16.5	17	50	16	67	16	25
Trautvetter et al. (2014)	6.5	0	6.5	19	59	30	48	30	24
Vieth et al. (2001)a ^a Trials for which habitual dietary intake was imputed from national survey data (age-, sex-specific).	5.4	25	30.4	33	43	17	65	17	–
Viljakainen et al. (2006b)	9.7	20	29.7	13	44	14	68	14	27.2
Viljakainen et al. (2006b)	10.6	10	20.6	11	47	10	61	10	25.8
Viljakainen et al. (2006b)	9.7	5	14.7	13	46	14	57	14	25.7
Viljakainen et al. (2006b)	10.9	0	10.9	12	52	20	44	20	25.6
Viljakainen et al. (2009)	8.6	20	28.6	16	62	14	90	14	24.4
Viljakainen et al. (2009)	7.6	10	17.6	16	60	12	76	12	24.9
Viljakainen et al. (2009)	6.6	0	6.6	16	65	19	52	19	24.8

BMI: body mass index; SD: standard deviation.
NB: e.g. (Madsen et al., 2013)a and (Madsen et al., 2013)b (as cited in Brouwer-Brolsma et al. (2016)) refer to the same study, but different population groups (e.g. in this case: children and adults).
^a Trials for which habitual dietary intake was imputed from national survey data (age-, sex-specific).
^b Values for Rich-Edwards et al. (2011) were imputed from Madsen et al. (2013) (children with same mean age).

Table 12. Risk of bias (RoB) dimensions – adequacy of randomisation, compliance assessment, dose check, overall RoB classification (35 trials)

Source	Randomisation adequate	Compliance assessed	Dose check	Overall risk of bias
Barger-Lux et al. (1998)	Yes	Yes	Yes	High
Barnes et al. (2006)	No/unknown	No/unknown	No/unknown	High
Bischoff-Ferrari et al. (2003)	Yes	Yes	No/unknown	Moderate
Bolton-Smith et al. (2007)	Yes	Yes	No/unknown	Moderate
Bonjour et al. (2013)	Yes	Yes	Yes	Moderate
Braam et al. (2003)	Yes	No/unknown	No/unknown	Moderate
Cashman et al. (2008)	Yes	Yes	Yes	Low
Cashman et al. (2009)	Yes	Yes	Yes	Low
Cashman et al. (2012)	Yes	Yes	Yes	Moderate
Cashman et al. (2014)	Yes	Yes	Yes	Low
de Gruijl and Pavel (2012)	Yes	Yes	No/unknown	High
DeLappe et al. (2006)	No/unknown	Yes	No/unknown	High
Forman et al. (2013)	Yes	Yes	No/unknown	High
Goussous et al. (2005)	No/unknown	Yes	No/unknown	High
Hansen et al. (2010)	No/unknown	No/unknown	No/unknown	High
Harris and Dawson-Hughes (2002)a	No/unknown	No/unknown	No/unknown	High
Harris and Dawson-Hughes (2002)b	No/unknown	No/unknown	No/unknown	High
Heaney et al. (2003b)	No/unknown	Yes	Yes	High
Heikkinen et al. (1998)	Yes	No/unknown	No/unknown	High
Holick et al. (2008)	No/unknown	Yes	Yes	High
Honkanen et al. (1990)a	No/unknown	No/unknown	No/unknown	High
Hower et al. (2013)	Yes	Yes	Yes	High
Keane et al. (1998)	No/unknown	No/unknown	Yes	High
Lehmann et al. (2013)	Yes	Yes	Yes	Low
Madsen et al. (2013)a	Yes	Yes	Yes	High
Madsen et al. (2013)b	Yes	Yes	Yes	High
Meier et al. (2004)	No/unknown	Yes	No/unknown	High
O'Connor et al. (2010)	No/unknown	Yes	No/unknown	High
Pekkarinen et al. (2010)	No/unknown	Yes	No/unknown	High
Rich-Edwards et al. (2011)	Yes	Yes	No/unknown	Moderate
Smith et al. (2009)	No/unknown	Yes	Yes	High
Trautvetter et al. (2014)	No/unknown	Yes	Yes	High
Vieth et al. (2001)	Yes	Yes	No/unknown	High
Viljakainen et al. (2006b)	No/unknown	No/unknown	No/unknown	High
Viljakainen et al. (2009)	No/unknown	Yes	Yes	High

e.g. (Madsen et al., 2013)a and (Madsen et al., 2013)b (as cited in Brouwer-Brolsma et al. (2016)) refer to the same study, but different population groups (e.g. in this case: children and adults).
Risk of bias dimensions and overall risk of bias as assessed by Brouwer-Brolsma et al. (2016).

C. Forest plots of achieved mean serum 25(OH)D concentrations by relevant factors explored in the dose–response models (random-effects meta-analyses) (35 trials, 83 arms)

Achieved mean serum 25(OH)D (and 95% CI) by RCT and sorted by intervention arm (n = 83 arms)

e.g. (Madsen et al., 2013)a and (Madsen et al., 2013)b (as cited in Brouwer-Brolsma et al. (2016)) refer to the same study, but different population groups (e.g. in this case: children and adults). CI: confidence interval; ES: estimate.

Weighted pooled estimates of achieved mean serum 25(OH)D by INTERVENTION ARM

CI: confidence interval; ES: estimate.

Weighted pooled estimates of achieved mean serum 25(OH)D by BASELINE MEAN serum 25(OH)D (nmol/L)

CI: confidence interval; ES: estimate.

Weighted pooled estimates of achieved mean serum 25(OH)D by TOTAL VITAMIN D INTAKE (μg/day)

CI: confidence interval; ES: estimate.

Weighted pooled estimates of achieved mean serum 25(OH)D by STUDY DURATION

CI: confidence interval; ES: estimate; mo: months.

Weighted pooled estimates of achieved mean serum 25(OH)D by LATITUDE

CI: confidence interval; ES: estimate.

Weighted pooled estimates of achieved mean serum 25(OH)D by STUDY START PERIOD

CI: confidence interval; ES: estimate.

Weighted pooled estimates of achieved mean serum 25(OH)D by ANALYTICAL METHOD

CI: confidence interval; CPBA: competitive protein binding assay; ELISA: enzyme-linked immunosorbent assay; ES: estimate; HPLC: high-performance liquid chromatography; LC-MS: liquid chromatography-mass spectroscopy; RIA:, radioimmunoassay

Weighted pooled estimates of achieved mean serum 25(OH)D by ETHNICITY

'African' was grouped to the 'Mixed' category, as this ethnicity included three arms only. CI: confidence interval; ES: estimate.

Weighted pooled estimates of achieved mean serum 25(OH)D by mean BMI of the study population

Normal weight: 18.5–24.9 kg/m², overweight: 25–29.9 kg/m², obese: 30 kg/m² and above. BMI: body mass index; CI: confidence interval; ES: estimate.

Weighted pooled estimates of achieved mean serum 25(OH)D by AGE

'Older adults': from trials where the reported or estimated mean age was ≥ 60 years. CI: confidence interval; ES: estimate.

Weighted pooled estimates of achieved mean serum 25(OH)D by SEX

'Both': for studies on mixed populations. CI: confidence interval; ES: estimate.

D. Forest plots of mean differences in achieved serum 25(OH)D concentrations (intervention arm versus control arm) by relevant factors explored in the dose–response models (30 trials)

Mean differences in achieved serum 25(OH)D by RCT (n = 30 trials) – random-effects meta-analysis

e.g. (Madsen et al., 2013)a and (Madsen et al., 2013)b (as cited in Brouwer-Brolsma et al. (2016)) refer to the same study, but different population groups (e.g. in this case: children and adults).

As indicated in Appendix C, mean differences in achieved mean serum 25(OH)D concentration were calculated for 30 RCTs, out of the final 35 studies included in the dose–response analysis, where a control/placebo group and at least one intervention group were available (i.e. 5 trials out of 35 did not have a control group).

CI: confidence interval; ES: estimate.

Weighted pooled estimates of mean differences in achieved serum 25(OH)D by TOTAL VITAMIN D INTAKE (n = 30 trials)

e.g. (Madsen et al., 2013)a and (Madsen et al., 2013)b (as cited in Brouwer-Brolsma et al. (2016)) refer to the same study, but different population groups (e.g. in this case: children and adults).

As indicated in Appendix C, mean differences in achieved mean serum 25(OH)D concentration were calculated for 30 RCTs, out of the final 35 studies included in the dose–response analysis, where a control/placebo group and at least one intervention group were available (i.e. 5 trials out of 35 did not have a control group).

CI: confidence interval; ES: estimate.

E. Model fitting

Table 13. Regression coefficients from meta-regression models as covariates are fitted (first row: null model; second row: ln of total vitamin D intake; last row: fully adjusted model) and related Tau², Adjusted R² and residual I² value changes

F. Meta-regression dose–response models; predicted mean serum 25(OH)D, 95% Confidence interval and 95% prediction interval

Meta-regression model of serum 25(OH)D response to ln of total vitamin D intake (unadjusted model) (n = 83 arms)

Circles represent mean achieved 25(OH)D in all included arms, either control or intervention arms. Mean achieved values are modelled against total vitamin D intake with no adjustments (i.e. no covariates) (Appendix C). The confidence interval (CI) illustrates the uncertainty about the position of the regression line. The prediction interval (PI) illustrates the uncertainty about the true mean effect that would be predicted in a future study.

G. Predicted achieved serum 25(OH)D and estimated total vitamin D intakes by AGE (adults, children) (74, 9 arms)

Table 14. Predicted achieved serum 25(OH)D (nmol/L) at selected values of total vitamin D intake (μg/day) by AGE

Table 15. Estimated vitamin D intakes (μg/day) at selected serum 25(OH)D cut-off values (nmol/L) by AGE

H. Sensitivity analyses

Table 16. Adjusted meta-regression models on subsets of the final data set after exclusions of trials with specific characteristics

Adjusted Ln of total vitamin D intake – μg/day (covariates coefficients not reported)	Coefficient	95% CI		Number of observations	Residual I-squared (%)
Final model	16.3	14.5	18.2	83	87
*Models restricted to trials without*
Recruitment of patient groups	16.4	14.4	18.4	78	87
Vitamin D supplement users	16.8	14.5	19.1	52	86
Persons with sun holiday during trial	18.0	14.9	21.2	41	85
Persons using sunbeds/artificial UV-B	16.5	13.3	19.8	31	78
Users of medication	16.0	13.8	18.1	42	85
Participants with diseases known to interfere with vitamin D metabolism	17.5	15.3	19.8	43	84

CI: confidence interval; UV: ultraviolet.
For detailed information on the concerned trials, see Brouwer-Brolsma et al. (2016).

I. Prospective observational studies

Table 17. Prospective observational studies – main study characteristics (n = 11 study groups)

Source	Publication year	Country	Latitude	Age Mean	Male %	Ethnicity (whole study)	Duration
Andersen et al. (2013)a	2013	DK	55.4	13	0	–	52
Andersen et al. (2013)b	2013	DK	55.4	72	0	–	52
Darling et al. ( 2013 )	2013	UK	51	34	0	Mixed	13
Darling et al. (2013)	2013	UK	51	38	0	Mixed	13
Hill et al. (2005)	2005	IE	51	60	0	–	52
Kift et al. (2013)	2013	UK	53.5	24	67	Asian	13
Lehtonen-Veromaa et al. (2008)	2008	FI	60.3	16	0	Caucasian	208
MacDonald et al. (2011 )	2011	UK	57	62	0	Mixed	65
MacDonald et al. (2011)	2011	UK	57	62	0	Mixed	65
MacDonald et al. (2011)	2011	UK	57	61	0	Mixed	65
Sullivan et al. (2005)	2005	USA	44	11	0	–	104

Source	Total vitamin D intake	Participants per group	Baseline Mean 25(OH)D	Baseline 25(OH)D SD	Achieved Mean 25(OH)D	Achieved 25(OH)D SD
Andersen et al. (2013a)	3.9	54	23	14	30	13
Andersen et al. (2013b)	8.1	52	47	25	51	24
Darling et al. (2013)	2.6	80	45	18	53	24
Darling et al. (2013)	2.0	26	20	11	22	11
Hill et al. (2005)	5.8	47	55	28	69	35
Kift et al. (2013)	1.4	86	20	7	15	7
Lehtonen-Veromaa et al. (2008)	4.0	142	48	20	48	17
MacDonald et al. (2011)	3.6	308	32	14	33	14
MacDonald et al. (2011)	3.1	114	44	18	46	18
MacDonald et al. (2011)	2.0	28	24	12	25	12
Sullivan et al. (2005)	5.4	20	56	17	51	14

DK: Denmark; FI: Finland; IE: Ireland; SD: standard deviation; UK: United Kingdom; USA: United States of America.
Eight prospective observational studies from seven articles were included, representing 11 study groups: (Andersen et al., 2013)a and (Andersen et al., 2013)b, as cited in Brouwer-Brolsma et al. (2016), refer to the same article, but different study groups (e.g. in this case: children and adults); a Caucasian group versus a Asian group were studied in Darling et al. (2013); a Caucasian from one study centre versus a group of Caucasian and a group of Asian people in another study centre were studied in MacDonald et al. (2011) (Appendix C).

Achieved mean serum 25(OH)D (and 95% CI) by STUDY GROUP

e.g. (Andersen et al., 2013)a and (Andersen et al., 2013)b (as cited in Brouwer-Brolsma et al. (2016)) refer to the same article, but different study groups (e.g. in this case: children and adults). CI: confidence interval; ES: estimate.

Achieved mean serum 25(OH)D (and 95% CI) by AGE GROUP

e.g. (Andersen et al., 2013)a and (Andersen et al., 2013)b (as cited in Brouwer-Brolsma et al. (2016)) refer to the same article, but different study groups (e.g. in this case: children and adults). CI: confidence interval; ES: estimate.

Achieved mean serum 25(OH)D (and 95% CI) by BASELINE MEAN serum 25(OH)D

e.g. (Andersen et al., 2013)a and (Andersen et al., 2013)b (as cited in Brouwer-Brolsma et al. (2016)) refer to the same article, but different study groups (e.g. in this case: children and adults). CI: confidence interval; ES: estimate.

Achieved mean serum 25(OH)D (and 95% CI) by LATITUDE

e.g. (Andersen et al., 2013)a and (Andersen et al., 2013)b (as cited in Brouwer-Brolsma et al. (2016) refer to the same article, but different study groups (e.g. in this case: children and adults). CI: confidence interval; ES: estimate.

J. Funnel plots of mean differences in achieved serum 25(OH)D from 30 RCTs (studies included in the meta-analyses) and Egger's test for small-study effects

Funnel plot of mean differences and Egger's regression line

As indicated in Appendix C, mean differences in achieved mean serum 25(OH)D concentration were calculated for 30 RCTs, out of the final 35 studies included in the dose–response analysis, where a control/placebo group and at least one intervention group were available (i.e. 5 trials out of 35 did not have a control group).

CI: confidence interval; SE of MD: standard error of mean difference.

Funnel plot of mean differences by vitamin D dose categories

As indicated in Appendix C, mean differences in achieved mean serum 25(OH)D concentration were calculated for 30 RCTs, out of the final 35 studies included in the dose–response analysis, where a control/placebo group and at least one intervention group were available (i.e. 5 trials out of 35 did not have a control group). e.g. '5-' means 5 to < 10 μg/day. CI: confidence interval; SE of MD: standard error of mean difference.

Citing Literature

Solarium Vitamin D Rã¶Hren

Source: https://efsa.onlinelibrary.wiley.com/doi/full/10.2903/j.efsa.2016.4547

Advertisement

Solarium Vitamin D Rã¶Hren